Until recently, American universities and media showed little interest in global rankings. Last year, however, interest in Leiden Ranking surged when Zhejiang University overtook Harvard at the top of its default indicator (here, here, and here). Then, several articles claimed the rankings were invalid or biased, and that Chinese universities were not really as good as the rankings suggested (here, here, and here).

The latest edition of the Shanghai Rankings (ARWU) has now appeared. Several commentators have noted that American universities still dominate the top of the rankings. Among them is Michael Nietszel writing on the Forbes page:

“Despite a recent history of governmental intrusion, political pressure and public criticism, all of which has captured extensive media attention, U.S. universities continue to dominate the Academic Ranking of World Universities (ARWU). In the 2026 version of the rankings, American universities claimed eight of the top 10 spots, and 14 of the top 2o.”

There are implications here. The author praises US universities for their ability to withstand “government intrusion, political pressure, and public criticism.” Elsewhere, he has recounted the alleged Trumpian assault on higher education. So, we can expect that as US universities fall in future Leiden and Shanghai rankings, it will be blamed on the evil designs of the big bad orange wolf and his cishteroableistnormative cronies.

Meanwhile, US higher education will be praised for its ability to thrive and lead the world ahead of those upstarts in China, Korea, and Singapore, or the laggards in Britain and the European Union.

But the reality is that the fall has already happened. You can see it in Leiden Ranking and, if you know where to look, in the Shanghai rankings. The latter have six indicators. Fields and Nobel awards won by alumni and those won by faculty, articles in Nature and Science, highly cited researchers included in the lists published by Clarivate, publications in the Science Citation Index Expanded and the Social Sciences Citation Index, and the sum of all these divided by the number of faculty staff, provided by verified government agencies.

When the Shanghai Rankings began in 2003, these indicators were generally internally consistent. I recall a conference presentation showing a correlation of around 0.7 for all of them, suggesting they were measuring much the same thing.

Since then, however, the rankings have shifted, and the correlation between the alumni and award indicators and the publication indicator has shrunk. Between Alumni and Publications, it is .36, and between Awards and Publications it is .35. The Alumni metric goes back to alumni graduating in 1921-1930 and Awards to staff winning prizes or medals in 1931-1940. In contrast, the publications indicator refers only to articles indexed in 2025.

As a result, the publications indicator is now seriously out of sync with the alumni and award indicators.

So, in the 2026 rankings, here are the top ten for alumni:

1. Harvard University

2. University of Cambridge

3. Massachusetts Institute of Technology

4. University of California, Berkeley

5. Princeton University

6. University of Chicago

7. Columbia University

8. California Institute of Technology

9. Yale University

10. University of Oxford.

And here are the top ten for Publications:

1. Zhejiang University

2. Shanghai Jiao Tung University

3. Harvard University

4. Peking University

5. Tsinghua University

6. Sun Yat-sen University

7. Sichuan University

8. Fudan University

9. Central South University

10. xian jtu

Yes, I know this is just one year, and Web of Science may not be the absolute pinnacle of academic excellence. Also, it does not include the arts and humanities. But, for better or worse, it is the future, and the awards and alumni are the past.

One suspects that the agenda of Forbes and similar publications is to heap the blame for the coming fall of the American academy onto Trump and his probable successors, denying the reality that it began decades ago.

Subscribe now

.

The Cambridge Scandal

A recent scandal at the Faculty of Education at the University of Cambridge shows that the university’s shining reputation has been built on rotting foundations. It will be interesting to see whether that reputation survives over the coming months and years.

Cambridge, Ranking, and Reputation

Cambridge, like Oxford, is rated highly by the commercial global rankings, especially for reputation. Times Higher Education (THE) has it in third place overall jointly with Princeton and behind Oxford and MIT. In THE’s last stand alone reputation ranking – apparently discontinued because Middle Eastern Universities were “gaming” the system – Cambridge was ranked fourth equal in the world, behind Harvard, MIT, and Oxford. In the QS world rankings, Cambridge also does very well, and again it does brilliantly in the academic reputation and employer reputation metrics. In this year’s rankings, it came 6^th overall but was 3^rd for Academic Reputation and 5^th for Employer Reputation. In noticeable contrast, it was 121^st for Citations per Faculty, QS’s measure of research quality.

Looking at the subject rankings, Cambridge is 4th in the world for Education Studies according to THE and, drilling down a bit, 3^rd for the Teaching “pillar”, 5^th for Research Environment, but 40^th for Research Quality. Turning to Cambridge’s performance in the QS Education and Training subject rankings, it is 6th overall, and 3rd for Employer Review and 6th for Academic Review. But again, Cambridge does poorly for citations: it is 19^th for H-Index Citations and 67^th for Citations per Paper, behind Manchester Metropolitan University and Middlesex University. There is a consistent pattern of Cambridge’s reputation running well ahead of its research prowess, measured by citations.

Turning to subject rankings that rely on objective and publicly accessible data, the university’s performance is not so strong. Cambridge is 50^th in the US News Best Global Universities Education and Educational Research ranking, 48^th in the Shanghai Global Ranking of Academic Subjects Education, and 31^st in the Scimago Institution Rankings for Education.

Whether Cambridge’s reputation will survive the current plagiarism affair remains to be seen. If it does, that would surely call into question the validity of the reputation surveys produced by the rankers, already undermined by problems with universities forming voting syndicates.

The Irresistible Rise of Jason Arday

The Arday affair, now all over the media, is threatening to capsize the wobbly boat of the university’s reputation. Jason Arday was allegedly diagnosed as autistic at the age of three, unable to speak until he was 11, illiterate until the age of 18, or in some versions of the story barely literate. Somehow he was admitted to Liverpool John Moores University and got a PhD from that university while working as a Physical Education teacher. Then came university appointments at Roehampton, Glasgow (or maybe not), Durham, and finally a Professorship in the Sociology of Education at Cambridge. Not forgetting honorary degrees from Southampton Solent University, for which he wore a Saville Row suit, and Anglia Ruskin University. He was also a Visiting Research Fellow at Ohio State University, Adjunct Professor at Nelson Mandela University, Trustee of the Runneymede Trust plus Olympic torch bearer at the 2012 Olympics.

All of this was while working as a cleaner and a trading assistant at Sainsbury’s, playing professional-level football and snooker, raising millions for charity, appearing on several TV shows, and sitting on advisory boards, taking part in the famous Seven Up documentary, a claim just retracted, running dozens of marathons, some with a fractured leg. He also had time to be a husband and a father. Unsurprisingly, he claims to have barely slept for several years.

I don’t think he has appeared on Strictly Come Dancing, Britain’s Got Talent, University Challenge, or the Great British Bake Off, at least not yet.

There is a report that he had epilepsy and another that he had an operation for a brain tumour just before the viva for his PhD thesis. Having passed his viva, he is reported to have had a stroke that caused substantial memory loss about the dissertation.

The Final Act Begins

His career would be truly inspirational if it had even a smidgen of truth. It is, however, increasingly hard to find that smidgen anywhere.

Now the stories have begun to unravel. In May 2023, David Harris, an Emeritus Professor at Plymouth Marjon University, tried to contact Arday about similarities between an article Arday had written for Social Sciences and a publication by Anjum Menom et al. in BMJ Open and an issue with informed consent from the subjects of the study. Arday refused further contact, saying “[a]nything further from you will be considered bullying and harassment.”

As recounted in Retraction Watch, Martin Hammersley of the Open University wrote to Hilary Cremin, head of the Faculty of Education at Cambridge, and Rhys Morgan, Head of Policy, Governance, and Integrity. Cremin replied that the article in question had been taken down and that Arday had the “full support” of the faculty, and Morgan said that the alleged malpractice had occurred before he was employed by Cambridge, so it had nothing to do with them. Later the Social Sciences article and another in Educational Philosophy and Theory were substantially corrected.

Eventually the data reached THE, which for a moment returned to its journalistic roots and prepared to publish an expose. Alas, that was not to be. Arday or the university got Carter-Ruck, the top gladiators among libel lawyers, to persuade THE to drop the matter. Yesterday THE gave its perspective on what happened: “THE and its journalists serve the higher education sector in the UK and globally to the best of their ability, seeking to champion its interests but also acting as a critical friend that asks challenging questions. This is what we were seeking to do in this case, and what we will continue to do in future.” Well, at least they tried.

There things rested for a while. Then an ex-Cambridge researcher entered from stage right. Nathan Cofnas, an American philosopher of biology and ethics, has degrees from Oxford and Cambridge, had been awarded an early career fellowship at Cambridge, and was also a College Research Associate at Emmanuel College. He has written on the heritability of human traits and group differences in intelligence, and his association with the college was terminated because his writings could “reasonably be construed as being opposed to the college values of diversity equity and inclusion”. That was upheld by Cambridge County Court, which held that Cofnas’s writing constituted a protected philosophical belief, but the college was allowed to get rid of him because the expression of that belief was “objectively offensive”.

Cofnas, who now has a temporary part-time post-doctoral fellowship at the University of Ghent, wrote in his Substack newsletter that he had analysed Arday’s LJMU doctoral dissertation using the Copyleaks software and found massive copying from an earlier Brunel University dissertation by Paula Zwozdiak-Myers. It was all there in black and white, or rather green and yellow, raising a huge question about LJMU’s supposed investigation.

That broke the dam and a flood of articles appeared first in the conservative newspapers the Daily Mail and the Daily Telegraph, then in the Times and the Guardian, and even the New York Times.

And now there is a full Turnitin report. This cannot be considered a definitive analysis since large parts of it consist of academic boilerplate. But it suggests that there was a lot more plagiarised than just the Zwozdiak-Myers dissertation. It also found that someone had actually plagiarised Arday.

Incidentally, Arday’s dissertation was cited just four times. That is not unusual in the social sciences, but surely something a bit more influential is expected from an applicant for a post at Cambridge, or indeed from any pre-1992 institution.

The Guardian piece added some new claims by Arday. He alleged that he had been threatened at his university office by a masked man with a knife and that a pig’s head and a banana had been sent to his family’s home, none of which could be verified by the university security or the police.

Nonetheless, Arday has supporters. The university dismissed the charge, saying that there was a “vile campaign” against him. A petition by the Good Law Project claiming that he had been vindicated by LJMU and Cambridge has been posted. It was signed by the heads of three Cambridge colleges, and a collection of peers, MPs, activist lawyers, and a few professors, and inevitably the leader and deputy leader of the Green Party. The number of signers has reached 14,607, although one college head has removed her signature.

The petition reads: “Entirely false allegations of plagiarism were made against Professor Arday and referred to two universities. Investigations have found no evidence whatsoever of wrongdoing. The relevant academic journals also found no basis for the claims.

Yet despite these unequivocal findings, rightwing media have attempted to smear Professor Arday’s reputation and undermine his career. That should concern everyone who believes in academic freedom, fairness and due process.”

The public controversy has so far revolved around Professor Arday’s plagiarism. Some have also noted that even if we set the plagiarism aside, what remains is of very poor quality. It is poorly structured, riddled with errors, and largely devoid of evidence. Much of it reads like a diversity statement or admissions essay by someone without a proof reader and rash enough to turn off Chat GPT.

His work is based on the assumption that Blacks and ethnic minority academics are subject to subtle and overt racism and are thoroughly marginalised. The lack of self-awareness is interesting: it is hard to think of anyone in academia who is less marginalised.

But even if it was not plagiarised and even if it was of sufficient quality, Arday’s work appears desperately thin for a candidate for any sort of position at a Russell Group university, let alone a Cambridge professorship.

Arday and his supporters are quite right to suggest that British universities have often been indulgent of the sins of White males. A notable case is William O’Reilly, an associate professor in early modern history, who is reported to have copied more than 12 pages from undergraduate essays. O’Reilly told an investigating tribunal, “ I firmly believe that I copied the material in my planning of the article, intending to use it only in a placeholding way, serving to remind me of the structure I planned to follow.” The tribunal found that the plagiarism was negligent and not deliberate. Maybe Arday can try that excuse, but it might not work this time.

It is also true that there is a bit of schadenfreude in the comments coming from some quarters. However, Arday’s academic misdeeds still need investigating although this affair has implications well beyond the writings of a single professor. First, the practice of accommodating students with “learning disabilities” needs to be reviewed. The Arday case is beginning to look like another example of the medicalisation of mediocrity. In a previous post, I noted that claims to mental disabilities were increasingly common in elite British and American universities, and were often used to favour the already privileged. Elite universities on both sides of the Atlantic are full of students claiming various disorders that entitle them to accommodations. A few decades ago, many of them would probably have just been regarded as not very bright and would have gone to a less highly ranked institution or found white or blue-collar employment.

Was the diagnosis of autism completely valid? Here I am reminded of an interesting case concerning an American Professor with a doctorate and a well known university leader, author of books and articles, who, even with massive accommodations for dyslexia and four attempts, could not pass the New York State bar exam, which at the time had a pass rate of around 70%. In other words, someone who could be an acclaimed research and university leader could not get to practice law, even at entry level.

Then there is the question of the two GCSEs that Arday might have got. If, in fact, he was illiterate or barely literate, that raises questions about the validity of those exams. If he actually was not, then the claims that he was autistic should be revisited.

Why and how he was admitted to St Mary’s University Twickenham to do a degree validated by the University of Surrey, and how he managed to complete the course there should also be examined.

Moving on to the question of the doctorate, Arday has claimed that he only saw his supervisor, Phil Vickerman, once and that everything else was done by phone and email. At this point, a certain scepticism is due about anything he says, but this is a claim that should be investigated. In any case, one wonders if Vickerman was aware of plagiarism and just waved it through. Surely it was standard practice to run dissertations and theses through Turnitin?

We might also think about how universities recruit their staff. In an era when there is a massive surplus of highly talented PhDs it seems odd that LJMU should appoint someone who failed all his GCSEs to a professorial position and appointment as a pro-vice chancellor: Arday’s supervisor “Pro-Vice Chancellor (Student Experience) Professor Phil Vickerman. Phil failed every GCSE aged 16 but went on to find a passion for learning, teaching and education after attending college and supervised Jason during his time at LJMU.” [my emphasis]

Miss Snuffy to the Rescue?

Anyway, the Faculty of Education and the rest of the University of Cambridge need to be thoroughly investigated. Nothing less is required to restore public confidence. Perhaps Katherine Birbalsingh could be appointed as interim Vice-Chancellor or at least head of the Education Faculty.

Subscribe now

Subscribe now

Times Higher Education (THE) have published a new edition of their University Sustainability Rankings, now renamed Sustainability Impact Ratings, which evaluate universities according to their alignment with the UN’s Sustainable Development Goals (SDGs). THE are presenting this as a new departure, creating a community, the Sustainability Impact Network, dedicated to the cause of global sustainability.

Membership in the community does not come cheap. A basic subscription is £3,620, which provides entry into the network. An advanced subscription, which gives “deeper data” on four SDGs, is £8,500 and a premium subscription for “institutions leading sustainability across their full portfolio and want clarity across all SDGs, access to historical data and dedicated support” costs £18,300. If that is still not enough, then there is an “Impact Ratings Enhanced Brand Package.” There are discounts for developing economies.

THE says that they will eventually stop publishing ranks and scores and classify institutions in bands or grades, but they are rather coy about when this will happen.

Although THE have presented their new “greater mission” with the usual fanfare it looks like a bit of a flop. According to THE, 1,646 universities have been evaluated this year, and results are displayed for 1,603, compared to 2,526 and 2,318 in 2025. That is a very significant fall, although THE claims to be satisfied, and it is possible that there will be more participants next year. That number is now less than the QS sustainability rankings, which include 2,021 institutions, although there is no option to withdraw from the QS rankings while participation in THE’s remains voluntary.

More significant is that the number of participants has fallen below those included in the Universitas Indonesia GreenMetric rankings, 1,745, something that has been ignored by university leaders and the media. Indeed, the mainstream media seem to have forgotten that GreenMetric has published sustainability rankings since 2010, long before THE or QS. Falling below GreenMetric ought to be an alarm call for THE, although their self-assurance is probably so great that they will not notice.

To briefly review the new rankings – it would be premature to call them ratings - the basic methodology is that universities are assessed, as before, on their compliance with the UN’s 17 Sustainable Development Goals. That assessment is based on a combination of research metrics and various data submitted by institutions or gleaned from public sources. Universities are ranked according to their three best scores and one mandatory indicator, Partnership for the Goals.

One result of this is that universities are encouraged to focus on just a few of the goals. If participation in these rankings has any relation to contributions to sustainability, then THE’s methodology will have the perverse effect of reducing the overall impact of many universities. Looking at this year’s results in detail, on average each university submitted data for ten goals, leaving seven unassessed.

These ratings now make explicit and intense use of AI:

“As part of our ongoing commitment to transparency and continuous improvement, we are incorporating large language model (LLM) technology into the validation of qualitative evidence submitted for the Impact Ratings. This section outlines how LLM is used, what it evaluates, and what universities should keep in mind when preparing their submissions. • All submitted evidence is reviewed and scored by our LLM model. The model determines whether each piece of evidence is specific, generic, or not relevant to the indicator question. • Although the LLM is the primary tool for validation, our human validation team remains in place to conduct quality assurance and review evidence as required. • If a piece of evidence cannot be processed by the model (for example, due to inaccessible formats), it is not automatically scored as not relevant. Instead, the system flags it and forwards it to human validators, who review and score it directly. • To ensure fairness, human validators follow the same visibility constraints as the AI—no clicking links or navigating beyond the visible content on a page. • Evidence must be self-contained. The reviewer (AI or human) does not click on links or explore additional pages to locate information. o Documents or web pages containing multiple links pose a challenge in ensuring fairness across institutions. In line with our policy of evaluating a single piece of evidence per indicator, such files are now being rejected and will no longer be accepted as valid evidence.”

I suspect that not everyone will find this reassuring. It is true that LLMs are becoming less error-ridden and delusional. However, it is unlikely that they will become completely error free any time soon, although they probably make the verification process more cost efficient. The rejection of documents with multiple links seems unduly onerous and will probably penalise many unsuspecting institutions.

Also, THE are now demanding one item of evidence for each data point instead of three. On the surface, that seems to reduce the burden on universities, but given the inaccuracy of AI and its human supervisors, it may well undermine the fairness and stability of the rankings.

Representation in the current edition is very uneven. There is minimal participation from China, with only three universities. Germany is also represented by three universities and the Netherlands, Belgium, and Israel by two each. Iran is totally absent this year and so is Russia.

In marked contrast, there is a substantial response from the Philippines with 160 universities, India with 110, Turkiye 74, Thailand 69, Indonesia 67, Pakistan 52, Egypt 45, Iraq 43, Malaysia 31, Japan 30, Saudi Arabia 28, Kazakhstan 24, and Bangladesh 22.

The turnout for the UK is quite good at 59 universities. For the USA, it is very sparse. There are just 16 American institutions, only one of them in California, and the highest ranked of these is MIT. The Ivy League, Oxbridge, the elite écoles of France and the Indian Institutes of Technology are all absent.

These rankings consist mainly of countries and universities that could not get anywhere in objective research-based rankings. They also fail an important test of validity, that is reasonable stability across successive iterations. Since the first edition of the THE sustainability rankings in 2019 there has been a massive turnover, partly due to methodological changes and partly to high-ranking institutions dropping out. Consider what happened to the top ten in the 2019 rankings. Table topper University of Auckland is barely in the top 100 in 2026, in 96^th place, McMaster University has gone from second to 18^th, King’s College London from 5^th to the 101-200 band, Bologna from 9^th also to 101-200. Five universities, British Columbia, Gothenberg, KTH Royal Institute of Technology, Université de Montréal, and the University of Hong Kong have left the rankings altogether. Only one 2019 top ten university is still there in 2026: Manchester, third in the first ranking, is now first this year.

This is extraordinary. For so many universities to rise or fall or disappear altogether casts serious doubt on the validity of the whole exercise. Comparing the results of the THE rankings with other sustainability rankings, of the top 11 in the THE rankings, exactly one can be found anywhere in the GreenMetric ranking, Universiti Sains Malaysia (USM). What about the QS sustainability rankings? There seems to be almost no relationship at all. Manchester is tenth in the QS sustainability rankings. The others are all much lower. USM is 247th, Griffith 76th, Queen’s University in Canada 59th, Western Sydney 172nd, Hanyang 162nd, Hokkaido 229th, Universiti Kebangsaan Malaysia 208th, National Taiwan University 72nd, Alberta 74th. Institut Agro in France is not there at all.

Now compare the top five UK universities with People and Planet, a ranking of the best British universities for sustainability which has been incorporated into the Good University Guide. Again, there is little connection. Manchester goes to 34^th, Glasgow to 96^th, Exeter to 28^th, Northumbria to 16^th, and Queen’s University Belfast to 50^th.

Looking from the other direction at the top five British universities in People and Planet. Manchester Metropolitan is 23^rd among British universities in the THE rankings, Bath Spa is 58^th, next to the bottom, and the University of West London is 50th; Reading is not ranked at all by THE, and Salford is 50th.

No doubt, THE would assert that their methodology is the best, even if it changes a lot, and the others are inferior.

A major problem with these rankings is that THE is looking at things that can be measured and may have little or no relationship with the stated goals. So, for example, SDG 16 Peace, Justice and Strong Institutions is measured by graduates in Law and Civil Enforcement, No Poverty by the proportion of students receiving financial aid, Zero Hunger by graduates in food sustainability, Good Health and Well Being by graduates in Health, Quality Education by the proportion of first-generation students, and Gender Equality by “Non-discrimination policies for transgender”.

A particularly ironic criterion is the measurement of Reduced Inequalities by the proportion of staff and students with disabilities. In the USA and the UK massive numbers of students are claiming accommodations such as extra exam time, lenient deadlines, and help with note taking, for a variety of conditions including depression, autism, anxiety and learning disabilities. It seems that the students who benefit from such privileges are disproportionately from wealthy families. It is irresponsible of THE to discourage such practices.

The new sustainability rankings may prove profitable for THE, but they are unlikely to be helpful for universities facing new and dangerous challenges over the coming decades.

The famous universities of the United Kingdom, especially Oxford, the country’s political-administrative class, and the ranking business led by Times Higher Education (THE) are engaged in a merry dance of delusion.

The commercial rankers proclaim that Oxford is the best, or nearly the best, university in the world, and that Cambridge and the London colleges, and maybe Manchester and Edinburgh, are top-30 material. Oxford and the rest of the higher education establishment defer to the rigour, robustness and sophistication of the rankings, and they concur that the upper civil service and political leadership, staffed disproportionately by Oxford and Cambridge graduates, are models of competence and efficiency.

The UK government has legitimised the rankings by declaring that they provide the membership of the Global Universities List, which is composed of institutions in the top 50 of at least two of the following: the THE World University Rankings, the QS (Quacquarelli Symonds) World University Rankings, and the Shanghai Academic Ranking of World Universities (ARWU). Graduates of these universities are deemed to be High Potential Individuals and can work in the UK without a sponsor.

Meanwhile, government spokesmen and establishment worthies have continued to praise the elite universities, sometimes citing their performance in the THE rankings. Back in 2007, Chris Patten, former MP, minister, Governor of Hong Kong, and Chancellor of Oxford, proclaimed that “another respected international survey placed Oxford and Cambridge joint second to Harvard in the league table of world-class universities. This confirms what others have suggested in recent years. Moreover, other British universities — most notably London’s Imperial College and University College — came out high on the list.”

Then in 2018 a parliamentary under-secretary of state claimed that “this is a time of significant and rapid change in our universities... our higher education sector is a national asset, attracting almost 450,000 students from outside the UK and consistently figuring in the top 200 institutions across the world—34 of those are British. Oxford University has been in the top position for the past two years in the Times Higher Education world rankings.”

The chorus of congratulations for British universities seems never-ending. In 2025, Rachel Reeves claimed that the “Oxford-Cambridge Growth Corridor is home to two of the world’s greatest universities and some of Britain’s most innovative businesses.” Bridget Phillipson said that they were “two of the finest universities in the world.”

Better than Harvard and Stanford, better than Zhejiang or Tsinghua? Look at the latest THE world rankings. Oxford is in first place overall. It is fourth for Teaching, second for Research Environment, and twelfth for Research Quality. In 2025 Oxford was second equal in THE’s last standalone reputation ranking. It is first for Computer Science and Medicine. Other UK universities also, according to THE, perform remarkably well: Cambridge third, Imperial College London eighth, University College London twenty-second, Edinburgh twenty ninth.

Phil Baty of THE was recorded in 2023 as saying that ‘Oxford University’s move to appoint its first female Vice Chancellor sent a “symbolic” wave around the world which created a positive image for the institution among academics.” That image apparently underlay Oxford’s remarkable success in THE’s reputation rankings.

The other “leading rankings” concur. The QS World University Rankings have Imperial, Oxford, and Cambridge in the top ten and the Shanghai ARWU has puts Oxford and Cambridge in that bracket.

The compliments are appropriately returned. Irene Tracey, Vice–Chancellor of Oxford, asserted that the THE world rankings are “rooted in internationally competitive research and teaching excellence, are recognised worldwide as a vigorous and demanding benchmark of university performance.”

According to the Vice-Chancellor of King’s College London, “[b]eing ranked amongst the top 2.5% of universities worldwide is a significant achievement and reflects the talent, dedication and ambition of our staff, students and alumni. Rankings such as these recognise academic excellence, but what matters most is the impact our community has through education, research and service to society.”

Meanwhile, Oxford prides itself on supplying the country’s political and administrative “elite”. Keir Starmer, Rishi Sunak, Liz Truss, Boris Johnson, Theresa May, David Cameron, Tony Blair, Margaret Thatcher and Harold Wilson went to Oxford. Over a third of senior civil servants are Oxford graduates, and two-thirds of permanent secretaries are from Oxford or Cambridge.

The Blavatnik School of Government at Oxford puts the British civil service at the top of its Index of Public Administration, in sixth place in the world overall, surpassed only by Singapore, Scandinavia and Canada, and third for public policy. Note that this index “explicitly excludes the measurement of policy outcomes (e.g. life expectancy, literacy rates, unemployment, economic growth etc), and wider aspects of public governance (the functioning of the legislature, judiciary, the rule of law, media/academic freedom etc).”

Of course it does. After all, they have to measure what really matters.

Praise for the civil service comes from many other quarters. Andrew Mitchell, a former cabinet minister, said “I believe the British Civil Service is one of the jewels in the crown in Britain”. Peter Hennessy, a Cambridge graduate, peer, journalist, historian and Kennedy Memorial Scholar, has talked about the “lustrous quality” of Britain’s “non-politically partisan public service, transferable from one Government to the next.”

That is the public perception. But reality is a little different.

Starting with Oxford, the THE rankings rely heavily on reputation, which gets a weight of 33%, divided between teaching and research surveys. Another 10% goes to income in 3 streams: institutional, research, and industry, with the biggest share going to research income. The Citations or Research Quality metric gets another 30%, much more than publications. That indicator was based entirely on Field Weighted Citation Impact (FWCI) until 2023, but since then it has been combined with three other citation metrics. FWCI is seriously biased towards books and conferences, where Western universities have a built-in advantage. There are more points for doctoral degrees awarded and doctoral students.

So basically, Oxford gets top marks from THE for being highly regarded by established researchers, having a lot of money, investing a lot in doctoral education, and being widely cited by other researchers, and, of course, by its own researchers.

One might also note that the THE rankings have done quite a bit to help Oxford. When THE introduced its long-overdue revamp of its Citations pillar, now renamed Research Quality, Oxford went from 28th to 5^th for that metric. In 2025, partly in response to unsporting behaviour by Middle Eastern universities, the reputation survey was enhanced with pairwise-comparison and voter-diversity metrics. In consequence, Oxford rose from 4^th to equal 2^nd.

Turning to rankings that focus on research or use objective public data, the latest Leiden Ranking, produced by the Centre for Science and Technology Studies at Leiden University, places Oxford 29th, down from 17th in 2015, for the default indicator, publications in core journals in the Web of Science. It is 74^th for Maths and Computer Science. In the Scimago Institution Rankings, produced by members of the Spanish National Research Council and Spanish universities and including data on Research, Innovation and Societal Impact, it ranks 26^th.

But the most catastrophic performance is in the highly selective and demanding Nature Index, based on a small basket of high-quality publications, where it is 32^nd. In Applied Science, it is 86^th.

From all sides, there is evidence that the abilities of Oxford students and the rest of the Russell Group are declining. In 2016, Oxford students began complaining that they were being made to work too hard and had too many essays. Some even reported working 50 or 60 hours a week. The horror, the horror! In 2017 Oxford history students were given take-home exams and maths students had extra time for their finals. The villain here appears to be the abolition of written entrance exams in 1995, which, combined with the degradation of A levels, forced colleges to rely on interviews, a notoriously inaccurate mode of assessment. The situation worsened as Oxford embarked on a campaign of social engineering that is reported to have seriously reduced academic standards.

In 2000 John Kay, former head of the Said Business School at Oxford, predicted that “Oxford University is set to lose its reputation as one of the world's finest academic institutions because of mediocre standards, outdated traditions and byzantine bureaucracy. Oxford’s reputation is, with a little help from its friends, intact for the time being, but in other significant respects it has entered a steady decline.

Last September, Harry Mount in the Spectator announced that Oxford had fallen in the Times/Sunday Times Good University Guide because of falling academic standards: “And so – no surprise – it’s just been announced that neither university was in the top three British universities for the first time since the records, produced by the Times, began 32 years ago.” Not quite. Oxford fell in those rankings because of a new metric in which it did not perform well, and the deletion of two in which it did. But declining standards, especially in the appointment and promotion of academic staff, have a great deal to do with Oxford’s fall in the rankings based on objective measures of research output and quality.

As for that other national treasure, the lived experience of anyone who has lived in the UK is that the civil service is incompetent, inefficient, overpaid, and far too fond of itself. Recently the Taxpayers Alliance reported that productivity has declined alarmingly despite massive pay increases. Pensions have ballooned so that senior pension pots are creating a class of mandarin millionaires. Falling standards are ubiquitous. The internal recruiting system has been described as “madness”. Former senior civil servants say that it is a “a system arrogant about its pedigree, pompous in its outlook, homogenous in its make up.” Recently it was announced that the Treasury had, in the name of diversity, dropped the test of numerical reasoning.

That would explain a lot.

Still, at least, our modern mandarins know what’s important. Consider Dame Antonia Rebecca Caroline Angharad Catherine Romeo DCB, Consul General in New York, permanent secretary in three large government departments, Westminster School, and then PPE, politics, philosophy, and economics at Oxford University, Starmer’s pick for cabinet secretary. She seems best known for an obsession with gender ideology: “Documents cited by The Telegraph suggest one employee was instructed to spend up to 20% of their working hours on “inclusivity” goals. These reportedly included promoting non-binary identities, encouraging colleagues to display preferred pronouns in email signatures, celebrating “non-binary awareness week,” and recruiting non-binary staff. Joining a “gender non-conforming book club”—where officials read novels centred on intersex and transgender themes—was allegedly written into performance targets that influenced career progression.”

And now we get to the rankings. Is THE really the gold standard of rankings? Not really. Maybe bronze. First, it lacks transparency. There are five “pillars” but 18 “carefully calibrated performance indicators”. It is impossible for anyone to figure out exactly what is going on except by paying for benchmarking or consultation packages.

These rankings have produced weird and whimsical results. For several years, the citation pillar, which contained a single indicator, produced interesting results. Universities in the top five for citations over the years have included Alexandria University, or more accurately a single serial self-citer and self-publisher, Tokyo Metropolitan University, Federico Santa Maria Technical University, St George’s University of London, Babol Noshirvani University of Technology, Brighton and Sussex Medical School, Anglia Ruskin University, and Cankaya University. As a result of methodological changes, this pillar looks a bit more sensible now, although the top 100 for supposed research quality still contains a few anomalies like Humanitas University, Australian Catholic University, Saveetha Institute of Medical and Technical Sciences, Prince Sultan University, and Graphic Era University.

Top of the charts for International Outlook are institutions in the smaller Gulf States, the SARs of Macau and Hong Kong, which THE treats as distinct states, Queen’s University Belfast, and Lincoln University Malaysia.

In 2021, a paper was published in Scholarly Assessment Reports that evaluated six leading global rankings according to four criteria: good governance, transparency, measuring what matters, and rigour. The authors refrained from assigning ranks or scores to avoid turning the study into another ranking, but the overall trend was clear enough. U-Multirank, now discontinued, and Leiden Ranking did relatively well, while “those rankings that are more highly relied upon by decision-makers, such as the Times Higher Education World University Ranking and the US News and World Report ranking, tended to score less well.”

The consequences of these mutually validating delusions are profound. Universities have failed to fulfil their primary purposes, the political-administrative class can at best offer a stylish mitigations

Subscribe now

of inexorable decay, and the commercial rankings continue to lose credibility.

A recent preprint by independent researcher Vladimir Moskovkin and a news story in University World News have highlighted the case of fake sites that mimic the Webometrics world rankings. There are now three sites with the domains .com, .org, and .online that incorporate webometrics in their title. The sites are attractive and look quite professional. Gone are the days, it seems, when you could tell a scam by shoddy presentation and poor English.

In several countries, including Spain, Portugal, Turkiye, Indonesia, and Montenegro, universities and the media have latched onto the fakes and published and celebrated the results. The real Webometrics rankings have announced they will exclude universities that cite the fakes.

Then there are rankings that cannot be definitively proven to be fake but provide no evidence of authenticity. First is Barentsz Uni-ranking. This also has an attractive interface, and the ranks and scores look plausible and unremarkable. MIT is first, then Harvard, Oxford, Stanford, and Berkeley. Tsinghua is 15th and Peking 19th.

Turning to the methodology, the site claims to have four metrics: Research, Reputation, Employment and Internationalisation. Supposedly, the rankings “rely on publicly available and third-party data to ensure objectivity and verifiability.” The source of the data is not specified. The page claims that the rankings are based on “the proprietary Uni-Ranking Data Engine, purpose-built by Barentsz DataLab to harvest, cleanse, and score vast information streams—producing the rigorous metrics that power our university rankings” and that “the Lab merges deep sector expertise with advanced quantitative techniques, delivering high-impact insights that shape strategic decisions and academic discourse alike.”

Sounds very impressive. There is more: “Barentsz DataLab is the analytics nucleus of our platform, trusted by dozens of Nasdaq-, NYSE-, and HKEX-listed corporations as well as leading financial institutions for objective, data-driven intelligence.”

So, what is this amazing data lab? Well, it seems that it was created by the Barentsz Group, which is “a Hong Kong–based consulting boutique that delivers strategy, management and technology consulting to publicly listed and growth-stage companies across Asia-Pacific”.

There is a website. The group claims to provide investment advice and has solutions, a mission, a story, and a vision, and adheres very passionately to DEI. The terms and conditions refer to Barentsz Capital Limited, which is registered in Hong Kong. It has an official legal existence, a market cap of 51.26M, several named executives, and 13 employees.

So there is a real company behind the rankings, but exactly how it arrives at the published rankings is unknown. Perhaps they have genuinely been scouring public data from OpenAlex and other public databases. Perhaps they have just added up the scores from the well-known rankings. Perhaps they have simply used AI to generate interesting-looking tables. Who knows? But until there is evidence, we cannot assume that these rankings are what they claim to be.

Another suspicious ranking is Uniranks. This is not UniRank, formerly known as IC4U, a legitimate search engine, which includes a ranking of universities by web activity. I suspect the similarity is not unintentional.

Once again there is a well designed page. There is a very detailed methodology for “A Student-Centric, 360-Degree University Evaluation System Empowering Students with Data-Driven University Insights”. For example, Internationalisation and Diversity has six indicators: International Students and Faculty, Global Partnerships and Collaborations, Student Faculty and Mobility, International Student Support, Multicultural Campus Environment, and International Reputation and Recognition. We are told that these metrics “ create a holistic assessment of a university’s global footprint, inclusivity, and ability to foster a diverse academic environment. Institutions excelling in these areas not only attract international talent but also contribute significantly to global knowledge exchange and cross-cultural understanding.”

Sounds good. But there is no indication of where this vast amount of data came from or how it is processed. In fact, there is nothing about the background of the rankings at all.

There is a LinkedIn page that proclaims that Uniranks is:

“the largest ranking engine ever made is a contemporary, unique, and innovative approach to the field of higher education, UNIRANKS is made in such a way that makes it possible for everyone to contribute and to be a part of this open-based project

UNIRANK engine utilizes 100s of ranking factors and 100s of millions of records for more than 30K universities around the globe and more than 2.5 million academics to achieve Accuracy and fairness.

This massive scale of data ensures that UNIRANKS platform will be unlike any other, Not only will it serve as a pivotal tool for promoting universities’ visibility but also will reflect their true capabilities.

UNIRANKS will serve as a one-stop shop for the students. Guiding them through the processes of the field of studies, Study destination, the Admission application process, internship, student life, and even career placement, and much much more.”

It seems there is a business address in New York City, and profiles for several “associate members” are listed. Meanwhile, the website links to a leadership certification seminar in Dubai starting next October, which has ten “advisory board” members and “program leaders”. One of these has already denied any connection with the event. I hope to report on the others later. There is no real information about how the rankings were constructed, so I think that this is just another bucket of AIslop.

Next is HE Higher Education Ranking. The claimed methodology of this ranking is that institutions fill out a questionnaire for which points are awarded. This year, the world leader is Sattan Astoni College London, UK, replacing last year’s champion, Charisma University in the Turks and Caicos Islands. The results are so improbable that the likeliest explanation is that scores and ranks are related to fees in some way.

The home page proclaims that it is listed as an “aspiring ranking” by the iREG inventory of international rankings. That was an error, and it was quickly corrected.

The advent of AI will mean that more and more “rankings” like these will appear. Everybody, especially universities and government agencies, needs to exercise extreme caution.

our trusted gateway to exploring the most has

Subscribe now

The advent of AI, the growing volume of available data, and increasing analytical sophistication are creating the space for new rankings and new assessment and consulting expertise. The latest rankings are from a London-based consulting firm, measuresHE, which is led by two former data experts from the Times Higher Education (THE) World University Rankings. It claims to offer insights and solutions to universities and publishes rankings that appear designed to bolster its reputation for analytical expertise. So far, it has produced rankings of journals and talented researchers, and it has now published rankings of national higher education systems similar to those from Universitas 21, discontinued in 2020, and QS in 2018 and 2020.

One wonders whether the long-term objective is to create a competitor to the THE and QS ranking-consulting-event management complex.

The firm says it has created a ranking that offers a “more holistic view of national research performance and global standing” and goes “beyond simple publication counts to reveal the true strength of academic ecosystems in the modern research landscape.” That sounds very persuasive and no doubt will impress university administrators, but it is perhaps a little imprecise.

Before digging into the methodology, here are the top ten countries according to the overall ranking:

1. United Kingdom

2. Netherlands

3. United States

4. Sweden

5. Canada

6. Australia

7. Germany

8. Switzerland

9. Hong Kong

10. Italy.

This is not very convincing. The evidence of the relative, and increasingly absolute, decline of British and American universities is so extensive and deep that this position of the UK and the USA in a research-oriented ranking seems counterintuitive.

Looking beyond the top ten, we can see that China is in 19th place, behind Finland and ahead of Saudi Arabia, Singapore is behind Spain, and Japan is behind Portugal and Ireland. That seems rather strange given the massive, well-documented increase in the quantity and quality of Chinese and East Asian research. It looks as though the ranking is presenting British and other English-speaking and Northern European universities as more capable than they really are.

These rankings are based on 25 metrics grouped into seven pillars. They are therefore more complex than the THE, QS or Shanghai global rankings. They are mainly, although not entirely, about research. There are metrics on the number of students and teachers, and tertiary education spending, but their weighting is relatively low.

The methodology is heavily biased toward countries with older, wealthier universities. Over a third of the total weight, 35%, goes to the Research “pillar”. Of that, a bit more than half, 18%, goes to Research Gravitas, which is meant to measure “a nation’s capacity to lead academic discourse”, and is determined by “a PageRank algorithm applied to subject-level citation networks, identifying structural influence within the academic community rather than mere citation volume.” I am willing to be corrected, but that sounds like the Research Influence metric in the THE World University Rankings, which “measures the level of thought leadership by considering how influential the citing papers are. It differs from the other FWCI-based measures as it examines not only the number of the citations, but also the “importance” of citing publications.” So, what this metric apparently measures is the extent to which researchers are cited by important researchers, whose importance is measured by their being cited by important researchers, which sounds a bit circular. The idea is to discourage universities boosting their ranking scores by soliciting undeserved citations. This would, however, create a self-referential system where it is difficult for outsiders to break into the closed circle of the traditional elite.

The next metric is Research Quality, which is “the outlier-trimmed arithmetic mean (Olympic mean) of the Field-Weighted Citation Impact (FWCI) for the country’s published works.” The Olympic mean, used for scoring at the Games, is calculated by deleting the highest and the lowest scores, thereby reducing the impact of outliers. This is a modified version of the THE Citation Impact metric.

Then we have Research Excellence, measured by “the proportion of a nation’s research works that rank within the global Top 5% based on FWCI [Field Weighted Citation Impact].” This is very similar to the THE Research Excellence metric, which is “calculated as the number of publications within the top 10 per cent of all publications by FWCI, normalised by year, subject, and staff size of the institution.” The threshold is moved up but otherwise it is essentially the same indicator.

The fourth research metric is authors in the Talent 100 rankings.

The second pillar, Sustainability, has a 7% weight. Five % is allocated to research gravitas, “a nation’s influence and thought leadership specifically within sustainability discourse. This applies a PageRank algorithm to SDG-related citation networks to identify foundational, highly impactful sustainability research.” Another 2% is for the proportion of research relevant to the UN Sustainable Development Goals. Countries get credit for doing research that is supposedly related to the SDGs, although a lot of those papers probably have a minimal relationship to sustainability.

Next, we have the Openness pillar, which is supposed to capture “the collaborative and accessible nature of a nation’s academic infrastructure.” To simplify, that means they look at the diversity of countries that a country collaborates with “by utilizing Shannon entropy to measure the ‘Diversity of Citing Institutions,’ the methodology goes beyond raw citation counts to reward true global distribution and penalize insular research networks.” Then there is industry collaboration, which means publications have industry partners, and open access research, which counts research published under open access licence.

This pillar is rather odd, since it includes three things that are quite different and are really united by the fact that the word ‘collaborative’ could plausibly be used to describe them.

The fourth pillar is International Integration, which is common in global rankings since everyone tries to track the nationality of their students, faculty, or research collaborators. Nothing really new here. It includes international research collaboration and international faculty and international students, which supposedly indicates “a country’s global appeal as a destination for emerging talent.” That is begging a rather big question, as it might also show a country’s need for cheap white-collar labour or ballot fodder.

Global Standing is the fifth pillar and carries a 20% weight. There is a 7.5% weight for the top two universities in the THE world rankings and 7.5% for those in the QS world rankings. Other world rankings, based in continental Europe and Asia, are not mentioned. Then there is 5% average subject-level Research Gravitas rank of the country’s top two institutions.

This would appear blatantly biased towards English-speaking countries. It is disappointing, since this could have been an interesting pillar had it included rankings from Spain, Russia, the Netherlands, Taiwan, and the UAE.

Then there is Demographics and Investment with a 10% weight, which consists of seven metrics: spending on higher education, number of students and teachers in higher education, number of researchers, gender equality for students and for teachers, and submission of data to UNESCO.

The final metric, academic integrity, has a weight of 10% and comprises penalties for retracted articles, author self-citations, and institutional self-citations.

Overall, there are some comments to make. The rankings show undue deference to the THE/QS duopoly, which are the only ones used for the global standing metric. If measuresHE were genuinely concerned about a truly global viewpoint, they would consider rankings from China (ARWU), the USA (US News), Spain (webometrics), Russia (MosIUR), the UAE (CWUR), or Taiwan (National Taiwan Rankings), rather than two UK-based companies. There is also the questionable globalist compliance with the UN, measured by their participation in the UN’s SDGs exercise and the punctuality with which they submit data to UNESCO.

Not only that, but research is measured by citations rather than publications. On the face of it, this would appear quite sensible. After all, it would seem that a paper that nobody or hardly anybody ever read would not be worth very much. That is true up to a point. But citations do follow publications, sometimes in a matter of years and in some fields in a matter of months, and so a metric based on citations will be less able to capture recent developments and it will in effect measure research capability from years or even decades ahead. We should remember that manipulating or soliciting citations is quite easy, sometimes virtually undetectable, and will become much easier as the AI revolution takes hold. In contrast, creating a fake article out of nothing is not so easy, although it can be done, as the Sokal or Boghossian and Lindsay hoaxes have shown. So a metric that does not take into account publications is seriously limited.

MeasuresHE has followed THE practice of lumping potentially very disparate indicators into pillars, which drastically increases the ranking’s opacity. However, they have, to their credit, provided Excel tables which can, when prodded, tell some interesting stories.

One thing that stands out is that the Research Integrity metric does not correlate very well with any of the others. The best is with International Integration, with a correlation of 0.36. The correlation with Research is very low at 0.14. So that means countries whose researchers cite themselves or their colleagues a lot, or whose publications are retracted, are less likely to be good at research, as measured by research metrics entirely based on citations. That seems reasonable. However, it does seem rather pointless to have a metric penalising author self-citations when it would be quite possible to exclude self-citations in the original count of citations.

These rankings are interesting since they are mostly based on the OpenAlex database, which represents a big step forward. However, they perpetuate the opacity and over-complexity of the THE rankings and fail to address the ongoing crisis in English-speaking universities.

Subscribe now

An interesting and puzzling article, ‘Mensa’s idea of ‘intelligence’ is plain stupid’, appeared in Spiked Online on the 4^th April. Michael Aaron Cody complained about Mensa, an organisation, perhaps club would be more accurate, that is restricted to those who score in the top 2% of people according to the Cattell III B test or its Culture Fair version or sometimes equivalent scores on the old versions of standardised tests like the SAT or GRE.

Cody, for some reason, seems upset about this rather pointless organisation. His article transitions from criticism of Mensa to a general attack on IQ testing and, by implication, standardised testing and public exams in general, with which it is, to varying degrees, correlated.

He is not alone. There is a medium-sized academic industry devoted to attacking the concept of intelligence testing. See here, here, and here for examples. He is scornful of IQ tests, which he claims have been discredited for decades. But IQ, or maybe more correctly, tests of general mental ability or cognitive skills, can in fact be considered one of the few success stories of the social sciences or at least of applied psychology.

The rest of the social sciences are going through a massively documented replicability crisis -- when you do the study again, you don’t get the same results, which is especially severe in education studies, where a succession of nostrums has come and gone. Mixed ability teaching, learning styles, discovery learning, multiple intelligences, self-esteem, whole books reading have, one after another, failed to replicate and have faded or are fading from classroom practice, although not so much from textbook dogmas and the curricula of education departments

In contrast, IQ studies -- if you’re triggered by this repulsive word, better stop reading now -- consistently return the same or very similar results year by year, and those results have direct and significant real-world applications. In most occupations, IQ is strongly associated with job performance more than anything else, except perhaps for expensive assessment centres. Nor is it just a matter of doing well in academic stuff, or the things that IQ measures. Life, as Linda Gottfredson has documented, is pretty much an IQ test.

Cody goes beyond the rather trivial but innocuous Mensa to assail the use of objective testing in academia and employment. His argument is based on rather flimsy evidence. Richard Feynman, Nobel laureate and certainly one of the brightest people who ever lived scored, according to Cody, only 125 on a school IQ test, which is bright but not spectacularly so. Well, not quite. That is what Feynman is reported to have said but in one version of the story it was his sister who found out when she is working in the school office and his IQ was even lower at 122, while hers was 123. This, frankly, is not at all convincing. Even if she had got into the school files and correctly identified the numbers that she saw there this is just one example of a test that was still in its infancy and probably very biased towards verbal IQ, which appears not to have been Feynman’s strong point.

Also, let’s remember that this was a century ago, and IQ tests have improved considerably since then.

Nonetheless, he has identified a significant element in current opposition to standardized psychometric and academic testing. Those who are clearly at the top of the heap for intelligence and would surely score extremely well on a properly administered test that they took seriously are very keen to announce that they don’t know their IQs or they never bothered to find out and anyway it doesn’t mean anything. It seems that ambitious and aggressive men and women take pride in high IQ scores or academic credentials when they are evidently more intelligent than their peers. But when they find themselves at an Ivy League school or Oxbridge or somewhere similar, they find that they are among people who are as bright or almost as bright as they are, they often react by claiming that the whole business of selection and testing is just a hoax or a swindle.

Cody, however, is riding the waves of recent history. Objective tests like the SAT, GRE, etc., have been relentlessly simplified and steadily stripped of their cognitive content so that they are increasingly tests of diligence rather than raw intelligence. Many American colleges have gone test-optional or have got rid of tests altogether, although there has been a retreat from that position since the fading of the COVID pandemic.

Similar things are happening to UK universities. Degree classes and A-level grades have inflated to the point that they are close to meaningless. Oxford abandoned its entrance exams in 1995 and has since adopted contextual admissions, further undermining the cognitive level of its students and graduates.

The consequences are dire. Harvard has to introduce what is in effect a remedial math course. Other elite schools have made remedial English composition compulsory for many students. Oxford students complain that they have to write too many essays. On both sides of the Atlantic, the graduate premium is eroding for many subjects.

In the last few decades, the leading institutions of the West have abandoned tests of cognitive ability or public exams with high cognitive content in favour of nebulous qualifications such as lived experience, judgment, resilience, responses to adversity, passion, and so on. The article complains that the tests do not account for brilliant strokes of genius. That is scarcely surprising, but the value of IQ is not that it will predict the next Darwin, Einstein, Mozart, Shakespeare, or Musk. It is that it can make reasonably accurate predictions about who will probably succeed as a talented but not miraculous researcher, or an efficient administrator, or a diligent and employable craftsperson.

And so, the research prowess of the universities of the West steadily declines compared to that of China, public services are visibly decaying, and everything is done poorly, if at all. Not to worry, it is possible that we might one day find a few geniuses if we keep looking for signs of neurodivergence, resilience, overcoming adversity, and so on.

Dynastic Decline

In his famous work Muqaddimah, the Arab scholar Ibn Khaldun described the process of dynastic rise and decline as lasting three to five generations. Similarly, the dynasties of ancient Egypt lasted an average of about 100 years, or 4 or 5 generations. The royal houses of medieval and early modern Europe typically lasted a bit longer.

And of course, we have the folk wisdom of Lancashire about clogs to clogs in three generations (leather shoes in the second).

Ibn Khaldun attributed this pattern to social factors based on the asabiyyah, which might be translated as social cohesion or solidarity, of dynastic founders and to the loss of such cohesion following the establishment of the dynasty and its corruption by decadence and affluence.

There could be a genetic element to this. It seems that most human attributes are heritable, at least to some extent, and that they are subject to regression to the mean, whereby the descendants of the elite, however that is constituted, fail to reach the standards of previous generations.

There are some political implications here. Choosing rulers by hereditary succession means they are likely to be better than anyone picked at random. But at the same time, the dynasty’s capabilities are likely to decline generation by generation. And so, in four generations, we go from Alfred the Great to Ethelred the Unready. If dynasties practice consanguinity or even incestuous marriage to keep wealth and power in the family, the decline is likely to be even more precipitous.

We can see this pattern in the history of the great financial and industrial families of the West. In the eighteenth and nineteenth centuries, the Rothschilds rose to dominate Europe’s financial system. Their intelligence, in both senses of the word, was undeniable. Among other things, they built a network of couriers to transmit and receive news before anyone else, although the story about a pigeon bringing the news of Waterloo to London is apparently apocryphal. The Rothschilds were also known for marrying their cousins, keeping everything within the family. But it seems that later generations are much less capable than the founders.

From the Epstein Files

Substacker Chris Brunet has drawn attention to an interesting section from the Epstein files. It seems that a young woman, an eighth-generation descendant of Rothschild founder Mayer Amschel, wanted, for some strange reason, to get into Columbia. Epstein apparently was influential there, but it seems that, even with the help of a Nobel Prize winner, there were limits to that influence, although one wonders whether he might have been more effective were he not such an obviously unsavoury character. But even a private tutor could not get her combined SAT scores over 1100, which is barely above average for all test takers. It is possible that the tutor prevented her from sinking below the mean, but it is certainly not the case that tutoring, test preparation scores, and all of the various social and economic advantages that a scion of the Rothschilds would enjoy could turn her into an Ivy-ready applicant. She did, however, get into New York University, ranked 28th in the most recent Shanghai Rankings, perhaps because she had written a dazzling admissions essay or excelled in some arcane extracurricular activity.

Similarly, crossing the Atlantic, we see that a young man with every possible advantage could only get a B in Art and a D in Geography in his UK A-level exams. According to The Jolly Heretic substack, there is now evidence that Prince Harry’s IQ is slightly above average. That, in view of his career trajectory, may be an overestimate, but it is nonetheless clear that all his advantages could get him nothing more than average grades.

The Campaign Against Testing

For many years, critics have denounced the use of standardised testing for admission to American colleges. It was claimed that they simply rewarded wealth because rich families could afford expensive test-prep courses or could provide extra tuition or coaching, and therefore such tests simply reinforce an unjust and corrupt social order. In any case, they did not test for real and valued attributes.

So, for example, the powerful organisation Fair Test asserts that:

“Such tests reward quick answers to superficial questions. They do not measure the ability to think deeply or creatively in any field. Their use encourages a narrowed curriculum, outdated methods of instruction, and harmful practices such as grade retention and tracking.”

The Intercultural Development Research Association (IDRA) makes a similar claim:

“Some students simply do not test well. Many students are affected by test anxiety or do not show their learning well on a standardized test, resulting in inaccurately lower scores.

Many students do not have a fair opportunity to learn the material on the test because they attend poorly-funded schools with large class sizes, too many teachers without subject area certification, and inadequate books, libraries, laboratories, computers and other facilities. High-stakes tests punish them for things they cannot control. If students do not have access to an adequate and equitable education, they end up being held accountable while the system is not.”

Across the USA, colleges and universities have accepted these claims and retreated from standardised testing. Many have gone test-optional, meaning prospective students can submit scores if they choose, and some have gone test-blind, refusing to look at applicants’ scores. Even before COVID, a string of reputable colleges had stopped requiring SAT/ACT scores for prospective undergraduates, the most prestigious of them being the University of Chicago, where some faculties also made the GRE optional for future graduate students.

During the pandemic, many more colleges went test-optional, and others, including the University of California system, went test-blind altogether. Although Harvard has now resumed testing, California remains a test free zone.

Back to Testing

The results have been decisively negative consequences. Harvard has had to introduce remedial math classes. Sorry, let me correct that: “a new introductory course aimed at rectifying a lack of foundational algebra skills among students.” The effects will reverberate throughout the system for a few more years, even if admissions policies are changed, and will likely erode Harvard’s reputation and status. The examples from the Houses of Rothschild and Windsor suggest that standardised tests and properly administered public exams are significantly resistant to gaming and manipulation, and that they are a necessary element in maintaining student quality.

It has become so common that it virtually qualifies as a genre: the stories of gifted students who have been rejected from a string of Ivy League and other elite colleges despite having perfect or near-perfect SAT scores, perfect or more than perfect GPAs, and participation in an array of sports and hobbies or passionate activism for approved causes. Equally predictable are the sneers from the commentators — he couldn’t write a proper essay, or had a defective personality, and so on.

Perhaps some American and Western universities are moving back to meritocratic policies. It does, however, seem that the anti-testing and holistic forces are very strong and in the end they may well prevail.

AI generated image

Ariel Procaccia, Alfred and Rebecca Lin Professor of Computer Science at Harvard University, has published a guest essay in the New York Times, ‘Don’t Trust the Rankings That Put China’s Universities on Top.’ It seems that he is leading a trend among American academics to reject the ugly reality of the rise of Chinese and Asian science.

Harvard, China and Leiden Ranking

His essay follows a spate of articles in the Chronicle of Highe

Subscribe now

r Education, New York Times, Forbes, Washington Free Beacon, Times of India, Xinhuanet, Global Times, and others that noted the rise of Chinese universities in the default indicator of Leiden Ranking, publications in the core journals of the Web of Science over a period of four years. Just two days before the time of writing, another essay by John Mac Glionn, ‘World University Rankings Are a Scam: China’s colleges are not superior to America’s’ was published by the James G Martin Center for Academic Renewal.

The articles focused narrowly on this single indicator for research in all subjects, which provides a useful although somewhat limited view of the decline of Harvard and other elite US institutions. However, Leiden Ranking publishes several metrics, including tables based on broad subject groups and the number of papers in the top 50%,10 %, 5%, and 1% of journals and these can supply a comprehensive and deeper view of developments in the research world.

Harvard is now third in the world for all publications, trailing Zhejiang and Shanghai Jiao Tong Universities but it is still first for publications in the top 10%, 5% and 1% of journals. So, that’s good for the USA? Not really. It is true that the higher up the quality ladder the better it is for American universities, especially for Harvard, but look again at the top 1% journal indicator, the top rung of Leiden’s quality ladder. For the period 2020-2023, Harvard is first, Tsinghua second, Stanford third, and Zhejiang fourth but fast backward to 2019-2022, and Tsinghua was third and for 2018-2021, it was fourth. Keep going back to 2006-2009 and Tsinghua was 116^th and Zhejiang 137^th. The pattern is clear. Chinese universities are catching up and overtaking those of the West year by year, month by month. The higher the quality, the longer it takes, but eventually, if current trends and policies continue, Zhejiang, Tsinghua, Peking, Fudan, and Shanghai Jiao Tong Universities and others, some scarcely known outside the PRC, will become dominant across the disciplines, especially in the harder sciences.

Performance in Subject Rankings

It should also be noted that Harvard’s position in the top ten or twenty of Leiden Ranking and other research-based rankings is largely derived from the strong performance of the Harvard Medical School and various affiliated hospitals. Look at the broad subject rankings for the Leiden Ranking default indicator.

For Life and Earth Sciences, the leader is the University of the Chinese Academy of Sciences, the best American University is the University of Florida in 7^th place, while Harvard is 73^rd.

For Physical Sciences and Engineering, the top spot goes to Tsinghua University, followed by thirty Chinese universities, and then the University of Tokyo in thirty first place. The leading US institution is MIT, in 53^rd place, and Harvard is128th.

The emerging Chinese superiority is equally clear when we look at the situation in Mathematics and Computer Sciences and check the total publications table for 2020-2023. The University of Electronic Science and Technology China is in the lead, and the top thirty-three places are all in Mainland China. I think it is worth being a bit irritating here: to repeat, thirty-three Chinese universities. Then we have Nanyang Technological University in Singapore in 34^th place, while MIT is 50th and Harvard 111^th.

But for Biomedical and Health Sciences Harvard still leads the world according to the quantity of research although Chinese schools now make up seven of the top ten.

Harvard Medical School Does the Heavy Lifting

Medical research by the Harvard Medical School now makes up the greater part of Harvard’s research output. In 2025, according to the Scopus database, Harvard published 34,670 research documents of all kinds, of which 22,504 came from Harvard Medical School. Slice the data another way and Harvard had 21,744 publications in Medicine compared to 2,865 in Physics and Astronomy and 1,916 in Computer Science. Compare that to Tsinghua University: 2,624 for Medicine, 5,471 for Physics and Astronomy, and 4,428 for Computer Science.

It is not too unkind an exaggeration to say that Harvard is now a very rich and very productive medical research center to which is attached a rather mediocre liberal arts college. Plus, of course, a massive and well managed pension fund.

To be fair, we should add that Harvard and other English-speaking universities still dominate the social sciences and the humanities and still maintain a precarious overall lead in Leiden Ranking’s Open Access Edition. For the moment.

Critiques of Chinese Science

No doubt some of those Chinese publications in physics, engineering, life and Earth sciences, mathematics, and computer science will be withdrawn because of errors or even fraudulence, many of them are mediocre or worse, and many are the products of a toxic work culture driven by ambition and greed. But given all that, it is clear that in these subjects China has caught up and is now well in the lead and the USA is falling behind. The denials of Professor Procaccia and others do not undermine the growth of Chinese, and perhaps some South and East Asian, continental European, and Middle Eastern, universities.

They note that until recently Chinese universities paid substantial bonuses to researchers who published in top journals, especially in Nature and Science, supposedly the peak of scientific research, and also that many papers from China have been retracted for one reason or another. It is true that China has officially ended the “pay for play’ research policy, although it surely continues in a low profile way through preferential admissions to doctoral programs, and access to grants, appointments and promotions.

There have been rushed projects and corners cut in pursuit of bonuses, but many Western academic administrators seem to think that researchers should work simply for the joy of it and cast their material ambitions aside, something that would be seen as condescending and insulting were it applied to athletes or musicians.

Comparing Chinese and American Universities

Sad to say, but in American universities, as elsewhere, you get what you, or your sponsors and benefactors, pay for. While China and some other Asian countries produce more and better research, America will get students, graduates, and faculty, who are passionate about social justice and sustainability, have overcome adversity, care deeply about their community, have certified “disabilities” and can write essays and statements that will impress mid-level bureaucrats or an upgraded LLM, or at least can find a professional how will help them “craft a narrative”.

It is standard practice for fading elites to search for ways to keep out upstarts from below and from abroad. The Ivy League and other US elite schools have developed a remarkably complex and opaque system of “holistic” admission, graduation, appointment, and promotion, primarily to restrict the inflow from unprotected and unwanted groups. Now we see the process repeating on a global scale, as the key players of the global higher education establishment, collaborating with the commercial rankers, scientific publishers and databases do their bit to deny the rise of China to scientific leadership.

The leading scientific database, Clarivate Analytics, has started a campaign to tighten up the requirements for its prestigious Highly Cited Researchers lists that has disproportionately affected Chinese, Indian, and Middle Eastern universities. The Times Higher Education rankings, popular with the media and university bureaucrats although derided by researchers, emphasise reputation and income and have revamped their rankings indicator to favour British and American universities while downgrading those in China. See the latest edition of the World University Rankings: Tsinghua University is 54^th in the world for Research Quality behind the University of Sharjah, the Australian Catholic University, Humanitas University, the University of Leicester, and Vita-Salute San Raffaele University. Not very convincing. QS Quacquarelli Symonds, the other leg of the global ranking duopoly, has now introduced a sustainability criterion into its world rankings, which has seriously penalised some Chinese institutions.

Harvard and the rest of the US elite may be better than China at some things and publishers, databases, the media, and no doubt ranking agencies will find new ways in which they can excel but that does not affect that China is doing more and better science and that is unlikely to change in the foreseeable future.

Procacia argues that Harvard does well at the things that really matter. For example, it has trained many Chinese PhDs who have gone on to be senior faculty at universities in China. Well yes, but that was a decade or two ago, and there is another way of looking at. Are American universities incapable of finding sufficiently qualified undergraduates from US colleges?

It is claimed that Washington under the Trump administration is the greater threat to American science and higher education. Maybe it is not helping very much but the crisis began decades ago, and it is very much a self- inflicted crisis.

The Academic Ranking of World Universities (ARWU), better known as the Shanghai Rankings, began with a clear purpose. It was to see how far Chinese universities had to go to achieve world-class status. The metrics are objective and public, and the path from raw data to published scores and ranks is fairly clear and transparent. They are, admittedly, limited and only measure research, saying nothing directly about teaching, sustainability, personal transformation, engagement with communities, upward mobility, international presence, or social justice. They also ignore the arts and humanities and underestimate the impact of the social sciences.

The methodology is quite straightforward. There are just six indicators: Nobel and Fields awards earned by alumni and faculty, papers in Nature and Science, publications in the Web of Science, highly cited researchers, and Productivity per Capita, which is the sum of scores for the above five metrics divided by the number of faculty. In comparison, the QS world rankings now have ten indicators, and THE’s have seventeen.

The initial rankings showed that China was still a long way behind the West, and that continental Europe was trailing the Anglo-Saxons. France was especially shocked by the poor showing of its universities and grandes écoles.

Since then, the rankings have maintained a generally stable methodology, apart from the inclusion of social science publications in the second edition in 2004 and fluctuations caused by Clarivate Plc’s changes to their lists of highly cited researchers.

Chinese universities have now substantially achieved the goal announced two decades ago. If we look at rankings that assess research based on publicly available data, such as the Leiden Ranking, Nature Index, or the Scimago Institutions Rankings, Chinese universities have reached, or even surpassed, the best American institutions. In mathematics, computer science, and physics, China is now significantly ahead, and that lead is likely to increase even further over the coming decades.

It is true that the QS world rankings have MIT at the top, and those published by Times Higher Education (THE) put Oxford in the lead, but a bit of shallow digging shows that they are there because of metrics like reputation, citations, income, or the number of graduate students. For criteria such as sustainability, open access, or gender equity, they perform poorly or are not competitive. A ranking recently published by TIME magazine did offer further hope for Oxford, which was placed at the top, but its validity is very questionable. For one thing, it did not include Zhejiang University in the top 500.

Returning to the Shanghai Rankings, Zhejiang University is now first for Publications, which measures articles published in a one-year period in journals in the science and social science indexes of the Web of Science. Shanghai Jiao Tong University is second, followed by Harvard. Unless there is a global catastrophe, more Chinese universities will surge past Harvard and the rest of the Ivy League over the next few years. For the Highly Cited Researchers indicator, based on lists curated by Clarivate, Harvard remains in first place, but Tsinghua University is third.

Harvard, however, still remains supreme in publications in Nature and Science and in faculty winning Nobel Prizes and Fields Medals. So, for publications and citations, China has caught up with the West. As for Nature and Science, it is questionable whether they now represent the apex of scientific research and it will take decades before China overtakes the West for Nobel and Fields awards, although that it is probably not worth bothering about.

For Shanghai Ranking it is, in effect, mission accomplished. ARWU identified the gaps separating the West and China, and over the last two decades China has closed those gaps and even gone beyond them in the significant metrics.

So the question now is what should China do next?

There are signs that China is preparing to assume leadership of the world of research and higher education. Recently, it rolled out the China Scholastic Competency Assessment (CSCA) for international undergraduate students, indicating preparation for a leadership role in global higher education.

It seems that there is now a chance for the Shanghai Rankings, or perhaps someone else, to take over the role of default world rankings currently held in many parts of the world by THE, and produce a ranking that would tell the rest of the world how far they are behind China and perhaps provide some idea about how to close the gap. To do that would require an overhaul of the current methodology while preserving the basic Shanghai principles of objectivity, reliability, and transparency.

There are signs that many universities around the world have had enough of THE and its lack of transparency and its arbitrary changes to an increasingly complex methodology. The Indian Institutes of Technology have boycotted their world rankings for several years, and major European institutions and Rhodes University in South Africa have recently joined them.

Meanwhile, the AI revolution is gathering pace, and it will soon be possible for rankers to process massive amounts of data, include many more universities, and publish rankings with monthly or even weekly updates.

So what kind of changes are needed? Starting with the Publications indicator, which measures articles and reviews over one year in the Web of Science Science Citation Index Expanded and the Social Science Citation Index, this could be made more inclusive by adding publications in the Arts and Humanities and including all the other research documents, such as proceedings, letters, and book chapters.

Shanghai have a point about excluding the humanities. In Western universities, there has been, in many respects, a noticeable decline in quality. But these disciplines do cover a substantial portion of the intellectual activity of modern universities, and it is likely that important insights into modern developments in sustainability, artificial intelligence, and biotechnology might welcome input from the humanities.

Something to consider is the bias these indexes have toward English-speaking researchers. This bias is frequently invoked in critiques of international rankings, although nobody seems to have any idea about a solution. If there is a bias, and if it is a serious flaw in the research system, the simplest thing would be to introduce a bonus for documents in languages other than English.

Next, papers in Nature and Science. It is debatable whether these journals can now be considered the apex of scientific research. They have become politicised and corrupted by current academic orthodoxies. The most outrageous departures from scientific objectivity were Nature’s 2020 endorsement of Biden for president because his “ trust in truth, evidence, science and democracy make him the only choice in the US election” and Science’s 2020 denunciation of Trump for “xenophobia, sexism, racism, transphobia, nationalism, and disregard for truth.” This indicator might be replaced by the journals used in Nature Index, or by the Top Journals and Top Conferences identified in the Shanghai Academic Excellence Survey.

The Highly Cited Researchers indicator has become problematic for Shanghai Ranking in recent years. Clarivate has started imposing stringent conditions on the inclusion of researchers, measures that have disproportionately impacted universities in China, South Asia, and the Middle East. Whatever the justification, it is surely anomalous that the assessment of global universities should be so dependent on a single Anglo-American company. It might be feasible to replace this indicator with a metric that would also include and give due weight to the Chinese Science Citation Database.

It might be time to replace the Award and Alumni indicators based on Nobel and Fields awards with one based on the Top Awards listed in the Shanghai Excellence Survey.

A frequent criticism of the Shanghai Rankings is that it focuses on theoretical research at the expense of practical applications. This might be amended to include patents and other forms of intellectual property, though the indicator should do more than just count filings.

Finally, the rise of AI is likely to render opinion surveys obsolete. Shanghai Ranking, unlike QS and THE, has wisely avoided such metrics. However, it should be possible to process large volumes of data to accurately assess public sentiment, something that Round University Ranking has already begun to do.

Western universities are facing a situation that China has confronted for more than two decades. Perhaps they will look at themselves honestly and take steps to reverse the current decline. Who knows, perhaps before long Yale or Cambridge may be known as the Tsinghua of the West.

Subscribe now

Subscribe now

Recently, there has been a big change in American attitudes towards global rankings. It seems to have started around last November when the Chronicle of Higher Education published an article highlighting the decline of elite American universities in Leiden Ranking and other global rankings based on public data and research excellence. A spate of articles followed, alerting readers to the USA’s dismal showing in these rankings.

Now TIME magazine, in collaboration with Statista, the German data analytics firm, has appeared on the scene with a new global ranking that shows that American, British, Canadian and Australian universities still dominate the global higher education scene, with German and Swiss institutions making a strong showing. In contrast, China is way down the charts. The top Chinese school is Tsinghua in 44^th place, just ahead of Brown, Dartmouth, and the University of Southern California. Zhejiang University, which is now leading in Leiden Ranking and the Publications indicator in the Shanghai Rankings, is not there at all. Neither are the Indian Institutes of Technology although the University of Delhi comes in 438^th.

The top university is Oxford, and the upper levels are dominated by American and British universities. After Oxford come Yale, Stanford, MIT, Chicago, Harvard, Cambridge, Imperial College London, Michigan, and Pennsylvania. Looking at the top 100 we can see 42 universities located in the USA, 19 in the UK, 8 in Australia, 6 in Switzerland, 5 in Canada, 3 each in Germany and Hong Kong, 2 each in Mainland China, South Korea, Singapore, and Japan, and one each in Denmark, Sweden, New Zealand, Ireland, Netherlands, and Austria.

These rankings appear to have been produced in a bit of rush. The absence of Zhejiang is surely an error since it meets the criteria for inclusion and has emerged as a stellar performer in the Leiden and Shanghai rankings. There is also an apparent duplication, Indiana University and Indiana University Indianapolis. Then, a couple of days after the initial publication, TIME announced that the tables had been revised due to changes to the Internationalisation indicators.

Turning to the methodology, there are three pillars. The first, Academic Capacity and Performance, has a weighting of 60% and includes resource expenditure per student, research income per faculty member, institutional income, faculty-student ratio, staff-student ratio, Nobel and Fields Laureates, citation counts, and highly cited researchers. Most of this pillar measures wealth and resources rather than achievement. Income is measured in three different ways, and human capital by the number of faculty and staff, which is likely to incentivise waste and inefficiency as universities will continue to boost their faculty and administrative numbers with one eye on the rankings. The rest of this pillar consists of measures of “high-quality research,” measured by prizes awarded for work done years or decades ago and by citations, but there is nothing about credit for recent papers. Looking at the ranks for this pillar, we have Oxford and Cambridge in first and second place, followed by Yale, Caltech, and Stanford.

Innovation and Economic Impact (30%) comprises patents, Massive Open Online Courses (MOOCS), traffic to university web domains, and top executives of companies listed on global stock exchanges. This is an incoherent assemblage that bundles technological progress, commitment to online learning, web activity, and alumni graduate careers in a limited sector of the world economy. Looking at the scores for this pillar, the leaders are Tsinghua University and Peking, Fudan, and Shanghai Jiao Tong, which score well, above Yale and Oxford.

Global Engagement (10%) covers international undergraduates, students, and faculty, as well as web traffic by international users. This pillar is dominated by British universities, with Scottish and Northern Irish institutions doing particularly well, a contrast with the THE and QS international indicators, where Hong Kong and Gulf universities usually take the lead.

TIME and Statista only provide the pillar weights, not the constituent metrics, which makes it difficult to analyse their results. I suspect that before long the data will become available for a modest price.

The weighting of the pillars is significant. If the weights for academic performance and innovation were switched, and that for global engagement was halved, then we have a very different looking ranking.

It looks as though TIME will follow US News and leave journalism behind, and become a processor and curator of data. There seems to be a ready market for rankings that will show that American and British universities are really the best in the world, and those newcomers in China, Korea, Southeast Asia and the Middle East are just chasing after quantity and cannot climb the heights where the Ivy League and Oxbridge preside.

Subscribe now

Until recently, the American political and educational elites ignored global university rankings. They were more concerned with the controversies about the national and graduate school rankings produced by US News. If they ever did consult global rankings, it was usually the Times Higher Education (THE) World University Rankings or perhaps the QS or US News global rankings. They hardly ever referred to the CWTS Leiden Ranking, URAP, published in Ankara, Webometrics, the National Taiwan University Rankings, or Round University Rankings, formerly based in Moscow, now in Georgia.

Now, suddenly, articles have appeared in the New York Times, Forbes, Ground News, the Washington Free Beacon, the Daily Beast, and Asia Times reporting that Harvard is declining, especially in Leiden Ranking, published by the Centre of Science and Technology Studies at Leiden University. It appears that the establishment has decided that the decline of American science and higher education cannot be denied, but it is only relative, and nothing to worry about, apart perhaps from Trump and his philistine policies.

Interesting, but this is old stuff. I have been reporting on Harvard’s decline in research-based global rankings for some time. Back in 2019, I predicted that Harvard would eventually be overtaken by Tsinghua University in Beijing in the publications indicator of the Shanghai rankings and estimated that this would happen in 2028. I have to admit I was very wrong. Tsinghua has not yet overtaken Harvard on that indicator. But Zhejiang University in Hangzhou did so last year.

By the way, Harvard lost its first place in another Shanghai indicator in 2022, when it was overtaken by Princeton for Nobel and Fields awards won by faculty.

I don’t want to seem immodest, but I suspect that my article in the Chronicle of Higher Education in November brought the US media to the realisation that something is wrong with Harvard, and with American science and higher education in general.

The New York Times has just published an article by Mark Arsenault about international rankings that references the latest edition of Leiden Ranking. This is unusual. Normally, rankings are assumed to refer to the US News rankings or sometimes the THE world or impact rankings. The article refers to Leiden Ranking’s default option, which counts publications in Clarivate Analytics’ Web of Science Core Collection, noting that on this basic metric, Harvard has now fallen below Zhejiang and Shanghai Jiao Tong University.

This is presented as news, although the Leiden ranking was published in October, and the story of Harvard’s decline is not new. I noted in the article in the Chronicle of Higher Education that Harvard and other American elite universities were falling behind, especially in engineering and computer science, and this decline was confirmed by looking at rankings based on public objective data rather than subjective surveys and institutional data. I discussed Harvard’s problems back in July 2022 and again in April 2025. Never mind, it is progress that the old grey lady, and perhaps some of the US establishment, recognises that Harvard is declining and China is rising.

The NYT article then cites Phil Baty, Global Affairs Officer of THE. On the surface, this is an odd choice, since the THE world rankings are, in terms of transparency, reliability, and consistent methodology, far from the best. Their complex structure is impenetrable to readers or stakeholders, and they have produced whimsical results for their international, citations and industry metrics: in 2022 they had An-Najah National University top of the world for citations, Asia University Taiwan for industry income, and Macau University of Science and Technology for International Outlook.

The article reports: ‘ “There is a risk of the trend continuing, and potential decline,” Mr. Baty said. “I use the word ‘decline’ very carefully. It’s not as if U.S. schools are getting demonstrably worse, it’s just the global competition: Other nations are making more rapid progress.” ’

In Asia Times Y Tony Yang says that, “Zhejiang University has surged to number one on the Leiden Rankings not because Harvard has faltered, but because Chinese institutions have dramatically expanded their research capacity.”

Not really. The hard reality is that US universities, and most European universities, have been in relative decline compared to China for some time, as measured by rankings based on objective research data and in the last few years they have entered the stage of absolute decline as measured by the number of high-quality publications in reputable journals.

Turning to the Leiden Ranking, one of its merits is that it provides the raw data on which the tables are based. The default indicator is the number of articles and reviews in core journals in the Web of Science Science’s Citation Index Expanded, Social Sciences Citation Index, and Arts & Humanities Index over a four-year period. In 2006-2009, Harvard’s count was 27,440. It rose to 36,732 in 2019-2022, although this was a slower rate than the Chinese schools. Then, in the period 2020 to 2023, it fell to 36,163, not very much, and maybe COVID had something to do with it, but definitely an absolute decline. It was the same for all the elite US universities: Yale, Berkeley, UCLA, Stanford, Princeton, Michigan, and Chicago produced less research over that period.

In contrast the number of publications produced by leading Chinese universities, Zhejiang, Shanghai Jiao Tong, Peking, Tsinghua, and even some that were barely noticed by the rest of the world like Sichuan and Central South University, have risen very substantially. In addition, the Indian Institutes of Technology at Kharagpur, Delhi, and Bombay and Middle Eastern universities like King Abdullah University of Science and Technology, Cairo University, and Qatar University have also risen.

OK, I hear journalists and rankers saying we are in the quality business, and you are just a philistine going on about quantity. So let’s take a look at the very highest level of research, the top 1% of publications. Harvard has fallen from 1,065 publications to 1,003 over four years, a decline of about 6%. Meanwhile, Chinese universities are rising and also some from India, Korea and Japan.

The problem is dire when we look at Physical Sciences and Engineering. In 2014-2017, Harvard was third in the world for high-quality publications in those disciplines with 145 articles and reviews, surpassed only by MIT and Stanford. By 2023, its count for those subjects had shrunk to 103, and it was in 24th place, behind 19 Chinese and 2 Singaporean universities, as well as MIT and Stanford. So we are talking about absolute decline.

The reasons for the decline are complex, but a major one is that, since the 1970s, American universities have lowered standards for undergraduate and graduate students and faculty. There has been massive discrimination against talented Asian- American students. Meanwhile, there is strong evidence that the selection and promotion of faculty is far from meritocratic.

What Steven Pinker said in 2014 is even truer today: “elite universities are nothing close to being meritocracies. We know that because they don’t admit most of their students on the basis of academic aptitude. And perhaps that’s what we should try next.” It is also true of the selection of faculty and administrators.

Subscribe now

The recent publication of regional rankings by QS (Quacquarelli Symonds) and Times Higher Education (THE) was received with scepticism in some countries and universities, reflecting negative perceptions of the commercial rankers.

A few years ago, it seemed that regional and speciality rankings were the future of university rankings. The rationale was that regions, countries, research communities, and disciplines had distinct priorities and were best served by rankings tailored to their specific needs. An additional benefit for the rankers was that regional rankings could be linked to lucrative summits or conferences at which rankers and leaders in higher education and research gathered to create and sustain networks and to improve and maintain their status.

A few universities, however, are disillusioned with rankings, particularly those published by commercial rankers such as THE and QS. In 2020, the leading Indian Institutes of Technology stopped submitting data to the THE rankings, citing a lack of transparency, excessive volatility, and reliance on reputation surveys. In 2023, South Korean universities, again citing a lack of transparency in addition to “mathematical flaws”, said they would boycott the QS rankings, although how they proposed to do that is unclear since the company normally replaces missing data with averages or previous submissions

Meanwhile, in September 2020, Sorbonne University joined Utrecht University, the University of Zurich, and Rhodes University in boycotting the THE rankings. The Sorbonne criticised the “black box” methodology, a bias toward English-language research, and an excessive reliance on reputation surveys. A few years ago, there was even a demand for a “boycott” of the Shanghai rankings in protest against the authoritarianism of the Chinese government. Leading American professional schools are less concerned with international rankings but have campaigned against the U.S. News law and medical school rankings.

Meanwhile, China seems to have lost interest in the Western rankings. It has largely ignored the THE University Impact Rankings—only eight participated in 2025—and the THE Interdisciplinary Science Rankings, in which only seven Chinese universities were represented in the top 200, fewer than those of Indonesia, Saudi Arabia, and Egypt. Many Chinese universities have received low scores in the QS Sustainability Rankings, particularly for environmental sustainability, suggesting that they are not making significant efforts. Another sign of disinterest is that the recent Shanghai conference on world-class universities featured only one speaker from Europe and one from North America, a significant contrast to previous years.

Much of this scepticism is directed at global rankings, but there has recently been growing concern about regional rankings. This has been accompanied by initiatives to create locally produced regional rankings, including the Applied HE rankings in Southeast Asia, with all-Asian aspirations, based in Singapore, the Arab ranking of Universities published by the Association of Arab Universities, and the EngiRank rankings of European technological universities, a product of the Perspektywy Education Foundation in Poland.

QS Asian Rankings

The latest edition of the QS Asian rankings appears to have been received with scepticism in China, with Alan Go noting some of these concerns in a LinkedIn post. There is concern that Fudan is ranked ahead of Tsinghua and Zhejiang, contradicting the conventional hierarchy and surprise that the top institution in Asia is the University of Hong Kong, in contrast to most other regional rankings. Also noted is the apparently strong performance of Malaysian universities.

The basic problem is that the QS rankings place considerable emphasis on reputation and international activity. The Asian rankings assign 30% weight to academic reputation, 20% to employer reputation, and 20% to international indicators. This works to the advantage of English-speaking universities in Singapore, Malaysia, and India, and to the disadvantage of those in Mainland China and Japan. Hong Kong universities gain an additional advantage because students and faculty from the mainland are classified as international in the THE and QS rankings.

The specific indicators in the QS Asian rankings reveal some counterintuitive results. For faculty-student ratio, we have Aga Khan University in Pakistan in first place, followed by Perdana University in Malaysia, and Shahed University in Iran. This seems a little suspicious, and one wonders whether the data from these places have been rigorously vetted. For International Students, the top ranked institution is Al-Madinah International University in Malaysia and for International Research Network, it is Islamic Azad University in Iran, followed by Zhejiang University and the University of the Chinese Academy of Sciences.

The research indicators also seem implausible. The university with the highest citations per paper, supposedly a sign of research quality, is the Asian University for Women in Bangladesh, followed by several Vietnamese and South Asian universities. Seventh for this indicator is the Singapore University of Technology and Design, which is ahead of the much better-known National University of Singapore and Nanyang Technological University. The first place for papers per faculty, a proxy for research quality, goes to Maulana Abdul Kalam Azad University of Technology in West Bengal. For staff with a PhD, the world leader is National Chiayi University, Taiwan.

The ranks for the other international indicators appear reasonable. City University of Hong Kong is first for international faculty, and inbound and outbound exchange students.

The THE Arab university rankings

The latest edition of the THE Arab University Rankings was published last November, and many of the results have caused some head-scratching. Some universities have soared dramatically while others have plummeted down the rankings. For example, Abu Dhabi University has risen from 14th to 7th, Lebanese American University from 12th to 9th, Prince Mohammed bin Fahd University from 24th to 11th, Prince Sattam bin Abdulaziz University from the 61-76 band to 16th, Gulf University of Science and Technology from 51-60 to 21st, and the American University of the Middle East from 38th to 26th.

Other universities have fallen significantly. King Abdulaziz University, which was first in the 2022 rankings, has gone from fifth to 15th, the University of Jordan from 9th to 33rd, Jordan University of Science and Technology from 21st to 69th, Cairo University from 8th to 45th, and the American University in Cairo from 13th to 45th. For some Iraqi universities, there has been a catastrophic decline. The University of Baghdad has gone from 20th to 151-175, Al Nahrain from 42nd to 176th, Mustansiriyah from 37th to 201+, Mosul from 41st to 201+, Anbar from 51-60 to 101-125, Babylon from 51-60 to 176-200, and Basrah from 51-60 to 201+.

In 2021, when the THE Arab University Rankings were launched, the methodology differed substantially from that of global university rankings. In particular, it included participation in the THE University Impact Rankings, which appears to be a dubious attempt to encourage universities to participate in rankings that were becoming financially unsustainable. This year THE cut back on some of its derivatives and realigned the Arab rankings so that they used the same indicators as the world rankings, although with slightly different weightings. Most of the universities that fell down the charts had previously performed well due to their participation and performance in the impact rankings. Some Arab universities rose to great heights that had nothing to do with merit, and then some crashed down the tables through no fault of their own.

It was therefore no surprise that several Jordanian universities announced that they would begin submitting data to the THE rankings. On October 7th 2025, the University of Jordan announced that it was withdrawing from the rankings due “concerns about the evolving methodology of the rankings and practices that affect the fairness and integrity of the evaluation process” and also because of the introduction of a paid subscription model for the THE impact rankings, soon to be ratings. On October 8th, two other Jordanian institutions, the Middle East University and Al al-Bayt University, also announced their withdrawal from the THE rankings. An article in the University World News reported that another three Jordanian universities had withdrawn.

The shift towards regional rankings is broadly welcome. It offers more specific and contextualised alternatives to the Procrustean global tables. It does, however, appear that the commercial rankers have overstretched themselves, and it is likely that more universities will decide to disengage from the rankings.

A big debate is brewing in the research world. There are problems everywhere with research misconduct, including plagiarism, data fabrication, and various shenanigans with authorship and affiliation, much of which is aimed at manipulating the rankings. The current centre of the storm appears to be India, particularly its private universities, where a significant controversy surrounds dubious publication, citation, and authorship practices.

Several websites and X.com accounts have catalogued the parade of dodgy, fake and plagiarised papers. For example, India Research Watchdog claims that Indian universities are now grappling with a research fraud crisis. Data show that retractions by Indian institutions increased 2.5 times between 2017-20 and 2020-2023. The reasons range from plagiarism to editorial conflicts to involvement in international research “paper mills.”

Integrity Index

Recently, Lokman I Meho, a medical researcher at the American University of Beirut, published an Integrity Index that measures data for questionable research practices. This has been updated, and the latest version has three indicators as follows.

• D Rate is the percentage of articles in journals recently removed “from Scopus or Web of Science for failing to meet quality or publishing standards”

• R rate refers to articles retracted because of “evidence of serious methodological, ethical, or authorship violations.”

• S Rate is “percentage of citations to an institution’s articles that originate from the same institution, emphasising citation practices aimed at artificially boosting institutional metrics rather than reflecting genuine scholarly influence.”

It is important to emphasise that these are indicators of potential malpractice, not conclusive evidence of outright fraud. Nonetheless, it seems they do highlight serious problems in the research cultures of many universities. It would be interesting to incorporate this kind of data into rankings or quality assessment.

Penalising Malpractice

In India, the National Institutional Ranking Framework (NIRF) has begun penalising universities for retractions. There have been suggestions that this approach could be applied to global rankings and that other indicators might be used.

According to Shuriah Niazi in University World News,

“Vineet Joshi, secretary of the Department of Higher Education, said: “Noteworthy changes in India Ranking 2025 included the introduction of negative marking for retracted articles. Moreover, the removal of self-citations under research and professional practices, applied across all categories and subject domains last year, was continued this year as well.”

Other countries, including Saudi Arabia and Egypt, have also implemented stringent measures against research malpractice.

This is a good idea, but there are some aspects that need consideration. A paper may be retracted for many reasons, for deliberate and conscious fraud or a technical issue related to the minutiae of authorship etiquette. Nevertheless, effective penalties for retraction could improve research quality.

But we still need to consider potential perverse consequences. If universities want to avoid penalties for retraction, they can, of course, stop doing the things that lead to retraction, which would be beneficial for everybody. They might also find various ways to circumvent the sanctions. Or perhaps they will become over-cautious, working on only safe topics, ploughing over lands that have long been fallow.

Self-Citations

A better approach is for rankings to encourage desirable behaviour as well as penalising bad practices. Referring to the S rate listed above, some progress has already been made. Leiden Ranking does not count self-citations, and the QS World Rankings have not done so since 2011. A further step, as the Meho report suggests, would be to bar citations from the same institution.

Excluding self-citations from the rankings would be a helpful first step. QS have been doing this since 2011, and it has always been an option in Leiden Ranking. Another step would be to exclude citations within institutions, as proposed by Meho, or at least assign them a reduced value. Another idea, which I suggested back in 2012, would be to exclude citations within journals, or perhaps even between journals with the same publisher. Universities would get more credit for citations from other institutions and other journals or publishers, encouraging a shift away from intellectual consanguinity. The emphasis should be on casting the citations net as widely as possible.

It is evident that Times Higher Education (THE) has been reluctant to exclude self-citations, even though this has led to highly embarrassing moments. or moments that should have been embarrassing. See here and here.

A Replication Indicator?

As noted in an earlier post, there is a massive crisis of replicability in the social sciences, especially in Education and Psychology. Studies often turn out to be deeply flawed in one way or another and cannot be replicated. A classic example is a 1966 paper in Psychological Reports by Rosenthal and Jacobson, which bloomed into a monograph entitled Pygmalion in the Classroom, that purported to show that students’ IQs were boosted just by telling the teachers that an IQ test had shown that they had high IQs and were about to blossom. The study was flawed and was not replicated. Nonetheless, it has become fodder for one of the great sacred cows of education, with an enormous influence, including the belief animating the ill-fated No Child Left Behind Act that every child would become proficient if only we rejected the “soft bigotry of low expectations”.

Perhaps we need a replication index that would give papers that are replicated a double or even a triple count in publication and citation indicators, and a similar negative score for citations of papers that fail to be retracted.

I may be wrong, but I suspect that this would have a greater impact on Harvard and Stanford than on mid-level universities in China or private colleges in India.

Once again, Oxford has proudly announced that it is the world's best university. The university’s home page reports:

“Professor Irene Tracey, CBE, FRS, FMedSci, Vice-Chancellor, University of Oxford, said: ‘We are honoured that Oxford has been ranked number one globally by Times Higher Education for the tenth year in a row. The THE rankings, rooted in internationally competitive research and teaching excellence, are recognised worldwide as a vigorous and demanding benchmark of university performance. This achievement reflects the dedication of our academics, professional services staff and students, but it comes at a time of real strain for UK higher education. Sustaining a dynamic and globally competitive sector requires renewed investment and support, so that universities can continue to drive discovery, opportunity and economic growth for future generations.’ ”

In fact, the THE World University Rankings are the only global rankings that put Oxford in first place. Looking at the well-known ones, Harvard is first in the US News Best Global Universities and the Shanghai Rankings, and MIT is first in the QS World University Rankings. In other rankings, including Webometrics, University Ranking by Academic Performance (URAP), the Center for World University Rankings, and the Moscow International University Rankings, Harvard is first, at least for the moment.

Meanwhile, Zhejiang University leads for the number of publications, the default indicator in the highly regarded Leiden Ranking, and the Chinese Academy of Sciences is first in Scimago Institutions Ranking and in Nature Index, arguably the most demanding of all the rankings.

So how does THE do it? What do they see that others do not? First, there is a significant weighting for teaching and research reputation, currently accounting for 33%, where Oxford, along with other big brand names, has a large inherited advantage. Not only that, but THE has recently made some adjustments that might benefit Oxford.

In February 2025, THE announced changes to the reputation survey, which now included indicators based on vote counts, pairwise comparisons, and voter diversity. The effect of these was probably to raise Oxford from fourth place in THE’s reputation rankings to equal second, overtaking Stanford and matching MIT.

Then we have citations. In the 2023 THE rankings, Oxford was ranked 27th. After some long overdue reforms in 2023, Oxford was eighth for Research Quality, as the citations indicator was now renamed, although it slipped back to 12th by the 2026 rankings.

In addition, THE has three indicators that measure income, institutional, research and industry. The only other international ranking that counts income is the Russian or ex-Russian Round University Rankings.

Then we come to Teaching Excellence. What does that mean? THE measures it with teaching reputation among researchers whose names are on Scopus publications, income per academic staff, doctorates awarded, and staff-student ratio. Sorry, but that does not mean excellence: it just means lots of researchers, lots of doctoral programmes, and lots of money, although perhaps not as much as Harvard or Yale.

THE is supposed to be a “vigorous and demanding benchmark”. Exactly what is a vigorous benchmark? I suspect that the author, whom ChatGPT assures me is human, meant rigorous. I doubt that THE rankings can seriously be described as rigorous. Most ranking experts, although not Russell Group PR officers, regard them with disdain.

THE is an outlier in putting Oxford at the top. When we look at the rankings or indicators that avoid surveys, focus on objective data, or use current information, then we find that Oxford is steadily declining. For research in the harder sciences, it is way behind the major Chinese universities, including some that until now are virtually unknown in the West, and sometimes Korean, Singaporean, or continental European universities.

In Leiden Ranking, to give just one example, it is 25th for overall publications, but 72nd for Physical Science and Engineering, trailing dozens of Mainland Chinese universities and a few Korean, Singaporean, and Japanese institutions, and even Sao Paulo, Purdue and the University of Tehran.

In URAP, Oxford ranks 40th for Information and Computing Sciences, behind the usual Chinese universities, as well as the University of Technology Sydney, the University of New South Wales, and the University of Sydney.

Of course, there is more to academic life than research. Maybe Oxford is best for something. Entrepreneurship? No, it is fiftieth in the PitchBook ranking of university start ups.

Gender Equity? No. In the Leiden Ranking, it is 585th for the proportion of female authorship.

Commitment to sustainability? Oxford isn’t even in the THE Impact Rankings or Universitas Indonesia GreenMetric world rankings.

So, sorry, Oxford is a long way from being number one in the world.

Subscribe now

The big news last week in the British higher education scene, or parts of it anyway, is that Oxford and Cambridge have fallen out of the top three in the Good University Guide published by the Times and the Sunday Times. This year, the top spot goes to the London School of Economics (LSE) for the second consecutive year, followed by St Andrews and Durham. Oxford and Cambridge were joint fourth, an unprecedentedly “poor” performance.

I was half expecting to hear laments about the lack of financial support from Scrooge-like governments, or the toxic fumes of Brexit, or maybe something to do with Trump. But so far, it seems that the current popular villain is the pressure to admit and to graduate students with limited academic potential.

According to an article in the Daily Mail, British universities “have been under pressure in recent years from successive governments to widen their intake to include more students from deprived backgrounds.”

Such policies have been criticised by Alan Smithers of the University of Buckingham, who said it was a “a very unfortunately political decision to insist that universities take into account social circumstances” and that “this will reflect on their apparent performance in the rankings.”

Professor Smithers is surely correct that the relaxation of admission standards at Oxford, Cambridge, and other UK universities has had a negative impact on British higher education. Combined with changes in the appointment and promotion of faculty, they have contributed to the steady decline of British universities in science and engineering in global rankings, such as the University Ranking by Academic Performance (URAP), the National Taiwan University Rankings, and Leiden Ranking, which are based on objective public data.

It is, however, rather difficult to see how this has anything to do with the decline in the Good University Guide league table. First, Oxford and Cambridge are not the only universities in the UK to embrace social engineering, and I see no indication that LSE, St Andrews, Durham, or anybody else has opted out.

Next, any time a ranking announces a significant or interesting rise or fall, it is wise to look at changes in methodology, such as the introduction of new indicators or adjusting the weighting of metrics.

The Guide now has eight indicators: Teaching Quality, Student Experience, Research Quality, Entry Standards, Graduate Prospects, Good Degrees (firsts/2:1s), Continuation Rate, and People and Planet, which has replaced Student-Staff Ratio in previous rankings.

People and Planet is a new indicator that purports to measure universities' contributions to sustainability, covering aspects such as carbon management, sustainable food, and ethical investment. I mentioned it in 2023 and noted that there was almost no overlap with the other sustainability-related rankings published by QS, Times Higher Education (THE), and Universitas Indonesia. Most of the universities at the top of this indicator are at the bottom for those that measure, or claim to measure, academic excellence.

Take the People and Planet criterion out of the equation, and Cambridge would be ahead of Durham and in third place. Bringing back the student-staff ratio metric would also be to the advantage of Oxford and Cambridge.

In addition, back in 2021, the rankings included Facilities and Spending, where Cambridge and Oxford did very well, along with LSE and Imperial College.

In contrast, Oxford and Cambridge have recently done badly in the teaching quality and student experience indicators, which are based on the National Student Survey. There was actually a boycott of this survey by students from Oxford and Cambridge a few years ago. Although the boycott appears to have ended, student experience scores remain mediocre. It is, however, unclear whether this is an objective assessment or just the result of disappointment that high expectations have not been met.

So, what happened is not the decline of standards at Oxford and Cambridge causing a fall in these rankings. I do not dispute that there has been a decline: that is confirmed, at least in STEM subjects, by the global rankings. They have fallen in the Guide because of a new and dubious indicator, the dropping of two favorable ones, and the unhappiness of Oxbridge students with their miserable experience.

Artificial Intelligence (AI) is spreading into all areas of the world economy. There is scattered but strong evidence that it contributes to significant improvements in worker productivity, although it still appears prone to errors and fabrication. Its impact is also evident in many aspects of research and higher education, and there is a widespread perception that it contributes to greater research productivity in preparing grant proposals, collecting and cleaning data, conducting literature searches, and revision and correction. It is likely that very soon we shall see entire papers written, reviewed, and read from start to finish by LLMs with modest human intervention, although their quality will probably be very poor.

It is possible that eventually the problem of AI hallucinations will be overcome or mitigated, and that we will see a new era of AI-driven discovery and innovation. There is, however, a risk that AI will simply reflect the biases and mediocre cognitive skills of its human trainers and auditors. Whatever happens, there is a good chance that the future belongs to those countries that make progress in the implementation and reform of AI and the selection of human intelligence to train and supervise it.

The following rankings are very simple: a count of all documents on the topic of Artificial Intelligence indexed in the Web of Science databases in 2024.

The first table shows that China has now overtaken the USA in the production of AI research. India is ranked third in the table, but if England, Scotland, Wales, and Northern Ireland were combined, the UK would be in third place.

Following those countries, the top twenty includes Western European, Anglophone, and East Asian countries such as Germany, Italy, Canada, and South Korea. An interesting feature of the table is that three Middle Eastern countries, Saudi Arabia, Turkiye, and Iran, are now in the top twenty.

The table for universities and research centres is headed by the Chinese Academy of Sciences, followed by the University of California system. Harvard is in fourth place, but it would be at the top if it included Harvard Medical School and medical affiliates.

Chinese, Asian, and Middle Eastern universities and research centers are performing well in the output of AI research. It is also noticeable that Saudi Arabian institutions are now forging ahead, with five Saudi universities in the top 200.

Subscribe now

The Shanghai rankings (ARWU) started over two decades ago. Their methodology is straightforward and simple compared to the complexity and opacity of the THE world rankings and their various offspring, but they can tell us a lot if you ask the right questions.

The latest edition of the rankings, released last week, shows at first sight a comforting story, one that might gladden the hearts of the American and European elites. Looking at the overall rankings, Harvard is in first place, as it has been since 2003, followed by Stanford and MIT. The UK is doing well with three institutions, Oxford, Cambridge, and University College London, in the top twenty. Continental Europe lags behind a bit, but its best performer, Paris-Saclay, comes in at thirteenth place. Then there is ETH Zurich holding up in 21st place.

Here we have a typical response from Université Paris-Saclay.

“On 15 August 2025, ShanghaiRanking Consultancy published its Academic Ranking of World Universities (ARWU). Ranked 13th in the world, and the leading university in France and Continental Europe, Université Paris-Saclay has once again confirmed its position as a world-class research intensive university.”

But the Shanghai rankings have in effect become two different rankings under the same label. The Alumni indicator counts Nobel and Fields award winners going back to the 1920s, and the Award indicator counts faculty winning those awards as far back as the 1930s. Unsurprisingly, these are a record of the research prowess of universities going back a century, although they are weighted towards recent years.

There is another indicator, PUB, that counts the number of articles indexed in the Science Citation Index Expanded and the Social Science Citation Index in the year preceding the publication of the rankings. For the latest rankings, that was 2024. There is, of course, a time lag between the start of a research project, the processing of grant proposals, the conduct of research, composition, submission, review, and so on, but this indicator is a reasonably current assessment of the research capabilities of universities.

It is, admittedly, not a perfect one since it does not take account of research in the humanities, nor does it consider innovation, teaching, progress towards sustainable development goals, or any of the amazing things that universities are supposed to do.

Looking a bit closer, we can see that the indicators are beginning to diverge. The correlation between alumni and publications for the whole dataset of 1000 universities is very modest, 0.36, and for Awards, it is 0.35.

However, when examining the top 200 universities in these rankings, no correlation is found between Alumni and Publications. The correlation for the first pair is 0.01, and for the second, it is -0.04.

Let’s pursue this point a little bit. There is effectively no link between the prizes and medals won by universities over the last century and their contemporary capability for scientific research.

Here we have a screenshot of the top twenty for Alumi in the 2025 Shanghai Rankings published last week. It looks very much like a snapshot of the research world around 1950 or 1960.

Then we have the top twenty for publications.

This looks like the research world as it might be in 2030 or 2040. Chinese schools, including many that most people in the West have never heard of, are in the lead. Harvard is still doing well, but I think I can safely predict that it will continue to decline for the foreseeable future. Then we have King Saud University coming in at number 10, and the University of Toronto, UCL, Oxford, Johns Hopkins, São Paulo, and Seoul National University.

As for the more distant future, it is wise, as Wayne Rooney is supposed to have said, not to make predictions, especially about the future, but it would be a good bet that there will be very few Western universities among the world’s research elite in another two or three decades. Except among Nobel and Fields laureates, if they are still in existence.

Subscribe now

Times Higher Education (THE) have announced that they will be using Artificial Intelligence (AI) to process the data submitted for their Impact rankings, now transformed into the Sustainability Impact Ratings. I suspect that will be followed by the use of AI in all sectors of the ranking business.

It would be naive to think that universities plugged into the THE ratings network will not use AI to produce the data required by THE. Media reports and the responses of university bureaucrats will likely be written with the help of AI. Ranking and other university-related activities will consist of massive amounts of discourse generated, transmitted, evaluated, and curated by a succession of Large Language Models (LLMs), without ever entering into a human mind, except possibly at the very end of the production chain.

Given the propensity of LLMs to hallucinate, that is to produce plausible-sounding answers that are wrong or nonsensical, this is unnerving. The future will be a world in which acquiescence in fictional or near-fictional narratives is an essential condition for employment.

Here is an example of what looks like an entirely AI-generated world university ranking. There is no information about data collection procedures, sources, advisory committee, or auditing, nothing to suggest that there is any relationship to reality.

On the other hand, the results do seem quite sensible: ETH Zurich first for research, Peking University for Reputation, University of California San Diego for employment. That is a bit more convincing than the THE world rankings, which in 2023 had Arak University of Medical Sciences first for citations, Istanbul Technical University for Industry Income, and Macau University of Science and Technology for international outlook.

So far, this ranking does not seem to be making any impact, but it seems there will be a new edition in September. Keep watching.

Subscribe now

This is a ranking of universities for women. It includes those higher education institutions that are reported as exclusively or primarily for female students and are included in the Ranking Web of Universities (Webometrics) published in January 2025 or any other global university ranking.

Institutions are listed in order of their position in Webometrics, which combines measures of research output with web activity. It is not a ranking of the best universities for women and does not say anything directly about the student experience at these universities, third mission activities, or quality of teaching.

We found that only two women’s universities not included in Webometrics were included in any other major ranking.

There are a number of girls’ or women’s colleges in India, and perhaps a few in other countries, that that are ranked by Webometrics. We have assumed that their exclusion is due to their not granting degrees, affiliation to universities, or lack of independent web activity. But please notify us if there are any errors.

Most women’s universities are in Asia, with the largest numbers located in India, Japan, and South Korea. However, American universities tend to cluster near the top, with American institutions taking six out of the top ten places.

In addition to world and country rank in Webometrics, we have also included ranks in the most edition of the world rankings published by Scimago, CWUR, THE, and QS. It is noticeable that Ewha Women’s does very well in these rankings.

The top ten universities are:

1. Ewha Women’s University, South Korea

2. Princess Nourah bint Abdulrahman University, Saudi Arabia

3. Wellesley College, USA

4. Smith College, USA

5. Barnard College, USA

6. Tokyo Women’s Medical University, Japan

7. Mount Holyoake College, USA

8. Bryn Mawr College, USA

9. Simmons College, USA

10. Alzahra University, USA.

Number of universities by country:

India 74

Japan 72

USA 27

Pakistan 19

South Korea 13

China 6

Bangladesh 2

Philippines 2

Saudi Arabia 2

Azerbaijan 1

Bahrain 1

Indonesia 1

Iran 1

Iraq 1

Kazakhstan 1

Kenya 1

Oman 1

Palestine 1

Peru 1

Sudan 1

UAE 1

Zimbabwe 1

World and Country Rank in Ranking Web of Universities January 2025